Good afternoon.

Welcome to today's New York City Council hearing for the Committee on Education.

Join with the Committee on Technology.

Please sign so all cell phone and electronic devices.

And as a friendly reminder, do not approach the days.

If you wish to testify today, please fill out an appearance card.

One of the sergeant at arms.

Without further ado, chairs, we are ready to begin.

Thank you.

Good afternoon, everyone.

I'm Council Member Eric Dinowitz, New York City Public School alum, although with a different signature on my diploma.

Public School Special Education Teacher and Chair of the Committee on Education, and I am excited to be joined by my friend from the South, Chair of the Committee on Technology, Council Member Delarosa for today's hearing on artificial intelligence, student data and privacy in New York City public schools.

Today's hearing will examine the landscape of AI use in New York City public schools, particularly as it relates to academic integrity and student data and privacy.

The rapid proliferation of AI tools has sparked intense citywide debate about the role of and use of AI in schools.

In May, DOE issued preliminary guidelines for AI in public schools, but the guidance has faced strong opposition from parents and advocacy groups alike.

Many council members, including myself, and numerous community education councils have gone as far as calling for a two-year moratorium on the use of student-facing AI in schools.

They point to ongoing worries about data privacy breaches and lack of transparency.

And very troublingly for me as a former teacher, cognitive offloading and algorithmic bias.

Others, however, argue that the question of allowing AI in schools is no longer relevant.

National survey data show that the majority of students are using some form of AI for school-related activities.

Instead, they argue the question should be: how do we harness it safely, equitably, and effectively for our students?

Underlying this debate is the issue of data collection by third-party vendors and whether DOE's policy aligns with national standards.

A 2026 audit by the state controller found that DOE does not have a documented data classification policy and does not maintain a comprehensive list of all application used by each school.

With a number of high-profile data breaches, including the power school breach that exposed personally identifiable information for tens of millions of students and educators nationwide.

Questions remain around what information is being collected about students and who has access to this information.

Now, as someone who doesn't even believe calculators should be used in elementary school classrooms, I look forward to hearing today about how DOE is planning to protect student data, plans for updated DOE guidelines and AI, and how DOE plans to ensure academic integrity and rigor in light of AI.

I'd now like to turn it to Chair Della Rosa for opening remarks.

Thank you so much.

Good afternoon, everyone, and welcome to our hearing.

Technology has become an unavoidable part of our lives, with artificial intelligence changing the way we work, learn, and make decisions.

This is why we have an obligation to ensure that innovation serves people and that technological progress is guided by accountability, transparency, safety, and public interest.

Artificial intelligence tools, artificial intelligence tools are used by the Department of Education in its daily operations, as well as in classrooms.

The use of AI in education has raised concerns that it may affect students' critical thinking and compromise the traditional learning process.

Another concern that people have related to digit is related to digital privacy, safety, and equality.

Some AI tools use personal data as training data.

That data could include students' information such as names, dates of birth, behavioral analytics, medical history, family income, ethnicity, and even facial geometry.

Without proper safeguards, this information could be used to profile students, predict or surveil their behavior, or make assumptions about their abilities and future opportunities.

This information can also be breached.

In fact, based on a recent New York State Comptroller's report, between January 2023 and February 2025, DOE experienced 141 data breaches incidents.

Based on our conversations and public reporting, some tools used by the DOE may collect data without parents' knowledge or consent.

While AI may offer potential benefits in education, its ethical use must be must be carefully considered.

Many tools may have bias and as a result discriminate against certain students or communities.

From the technology committee's perspective, safeguards against bias in AI algorithms and protecting student privacy are fundamental issues that must be addressed.

Today, we wish to discuss both how AI should fit into the curriculum and how the Department of Education uses AI in its operations.

We need to make sure that AI supports students and teachers rather than undermining the quality of education.

We must ensure that we are preparing the next generation to thrive in this technological era while also protecting the rights and future opportunities.

Like Chair Dinowitz, I also joined my colleagues in signing on to a letter from our school communities, and we've heard from our CCs calling on a two-year moratorium.

We look forward to hearing testimonies from the Department of Education and OTI to gain insight and clarity around the many questions about AI in the classroom and the use of AI in the education system.

We also look forward to hearing from members of the public, students, families, educators, and advocates.

Unfortunately, neither the chief privacy officer nor the representatives of the newly established Office of Algorithmic Accountability are here today.

However, we look forward to receiving their testimony, which we've received, and responses in writing.

As the school year is ending, I also wish to congratulate every student and teacher on completing another academic year.

Now I'd like to thank committee technology committee staff, legislative counselor, counsel, Irene Bybolski.

I'm I'm learning.

I'm gonna get it right, I promise.

And policy analyst Eric Brown for putting together this hearing.

I'd also like to recognize uh members of the council who are here of the technology committee, uh councilmember Shulman, Hanif, and Gutierrez, and we have Swang and Wang here as well.

Thank you for being here.

Am I missing anybody?

No?

Okay.

Um, and now we'll turn it back to Chair.

Thank you.

Thank you, Chair.

Additionally, today the committee will also consider the following legislation.

Two bills sponsored by Councilmember Susan Zhuang.

These include Introduction 684 in relation to requiring a study on how language access needs and geographic location affect access to early elementary gifted and talented programs.

An introduction 853 in relation to requiring a study on how language access needs, income, and geographic location affect access to special education programs and services.

I'd like to invite Councilmember Zhuang to make um opening remarks about these bills.

Thank you, Chair.

I'm here today asking for your support for my education access for all bill package, intro 684 and the intro 853.

I'm proud to put forward this pair of bills that together we would go a long way towards expanding education access for all students.

Every child deserves the education that meets them where they are and help them grow.

But like everything else in this city, children in other boroughs have less access to specialized programs than others.

Children from lower-income families have less access.

And the children for whom English is a second language has less access to.

I believe that all students deserve access to education programs that meet their needs.

And that no students should be limited due to where they live, what's their parents' income, or language they speak at home.

Both of these bills ask a city to do a study how language access needs, income, and the geographic location affect access to education.

Intro 684 is focused on access to gifted and talented program, while intro 853 is about access to spatial education program.

Gaps in access affects both types of specialized education.

For example, in sadly, well known, not enough black and the brown students in specialized high schools.

But how can we how can brilliant black and the brown students have the same chance to get in when their elementary school and the junior high school do not have gifted and talented program that other students have?

We are denying black and the brown students opportunities, and that must be it.

We should not be thinking about eliminate those programs.

We should be expanding them so every child who can benefit from them can get in to one.

We need to create more opportunity, not less opportunity.

As another example, it's also sadly well known.

Some children in outer boroughs need to travel alone bus rides to get to the school, offer the special education they need.

It's a extremely burden for both students and their parents, especially in the bad weather.

I still remember that parents call my office in the snow day.

The school is open, they have to travel, the kids have to travel more than two hours to get into school, and spend one hour, two hours later, and this school have to end early, and then travel another two hours to get home.

Parents, those first language is not English, also struggle to learn about opportunities available to their children and how to get their children into those programs.

These are only a few examples.

And I also want to thank all of you who have already co-sponsored and advocates came to testify and rallied with me today.

Thank you.

Thank you, Councilmember.

Finally, the committee will hear two bills sponsored by myself.

Pre-considered intro 2129 in relation to annual reporting on algorithmic tools approved for use in schools, and pre-considered introduction 2130 in relation to quarterly reporting on data breaches of algorithmic tools in schools.

As we will discuss today, AI is already entering schools in ways that are meaningfully and perhaps detrimentally shaping the educational experiences of students, as well as the professional experiences of teachers, counselors, administrators, and everyone in between.

That makes transparency critical.

Families and educators should know what tools are being used, what data they collect, whether they have been tested for bias and accuracy, and what happens when sensitive student data is exposed.

By requiring regular public reporting on approved AI tools and related data breaches, these bills will help turn AI oversight from a matter of trust into a matter of accountability.

I look forward to hearing testimony from the administration about these pieces of legislation.

Thank you to all the members who are here that have been named by Chair De La Rosa.

I would also like to thank my chief of staff, Jenna Klaus, and my legislative and budget director, Theo Salter, the education committee staff, including Council Alejandro Carvajal, senior policy analyst Chloe Rivera, policy analyst Katie Selena, Principal Financial Analyst Andrew Lane Lawless, and financial analyst Grace Amato for their work on today's hearing.

I would like to remind everyone who wishes to testify in person today that you must fill out a witness slip, which you can find on the desk of the sergeant at arms near the entrance of this room.

Please fill out the slip even if you have already registered in advance that you will be testifying in person today.

I also want to point out that we will not be voting on any legislation today.

To allow as many people as possible to testify, testimony will be limited to two minutes per person, whether you are testifying in person or on Zoom.

I am also going to ask my colleagues to limit their questions and comments to five minutes.

Please note that generally, witnesses who are here in person will testify before those who are signed into the Zoom webinar.

I would also like to note that we have been joined by council members Linda Lee and Kevin Riley.

Okay.

As well as Councilmember Fair Lewis.

As a reminder to all of our witnesses, please state your name prior to your testimony for the record.

If anyone here today if anyone here today requires an accessible version of the presentation given at today's hearing, please email testimony at council.nyc.gov.

Now, in accordance with the rules of the council, I will administer the affirmation to the witnesses from the mayoral administration.

I will call on each of you individually for a response.

Please raise your right hand.

There's a lot, a lot more papers than you're up here.

Do you affirm to tell the truth?

Please raise your right hands.

Thank you.

Do you affirm to tell the truth, the whole truth, and nothing but the truth before these committees and to respond honestly to council member questions?

Alicia Osborne Reinhardt.

Thank you.

Jeff Klein, Veronica Queller, Dennis Doyle, Tara Craza Croza.

Got it.

Okay.

Lara Lai.

Brenda Garcia.

Suzanne Sanchez.

Liana Wright.

Scott Strickland.

Alan Chang.

Maribel Hulla.

FDC Danielle Ginta.

Mia Teresa Pate.

Pate.

Thank you.

If you're here, is tweet empty?

You're all here.

Alright, thank you.

You may begin.

You may begin your testimony.

Thank you.

Thank you, Chair Dinowitz, Chair De La Rosa, and members of the committees of education and technology.

Good morning to our council members.

My name is Danielle Junta.

I am the first deputy chancellor of New York City Public Schools serving under the Chancellor of Kamar Samuels.

Prior to this role, I was a teacher, network leader, superintendent, and most recently uh chief of school leadership development.

But my favorite role in New York City Public Schools was as a student.

I am a graduate of New York City Public Schools.

I now oversee superintendents and key offices, including academics and instruction, district planning, enrollment, new school development and design, policy and evaluation, and student pathways.

I'm also joined today by supervising superintendents, Dr.

Alan Chang and Dr.

Maribel Hullah who work directly with our superintendents and their school communities across the city.

And because AI spans multiple divisions and areas of expertise, I am also joined by colleagues from data privacy, the division of instructional and information technology, early childhood education, and the division of inclusive and accessible learning, underscoring our commitment to a coordinated and responsible approach to this work.

Over the past few months, the chancellor and I have been speaking regularly to our communities about artificial intelligence and screen time in classrooms.

Those conversations have made clear that families, educators, and students are thoughtfully considering how these developments are shaping teaching and learning for our students.

I appreciate the opportunity to gauge with you about artificial intelligence, student data privacy, and the role of technology in our schools.

We know this conversation is urgent.

We recognize the rapid evolution of AI requires careful judgment, clear safeguards, and a meaningful dialogue with communities we serve.

Families, advocates, labor partners, school communities, and members of this council have pushed us to think more deeply about the role of AI and screen-based technology in our schools.

We are grateful for the engagement, including the lender letter sent to the mayor and chancellor by 29 members of this council.

The concern raised by the council members and school communities have helped shape our approach and reinforce the need for clear guardrails, transparency, and ongoing engagement.

When Chancellor Samuels assumed his role in January, New York City Public Schools was completing AI guidance that had been initiated under the prior administration.

On March 25, New York City Public Schools published that guidance, launching an AI landing page on our website, opened a 45-day feedback window, and began collecting input from educators, families, students, and community members across the city.

This process was designed to be iterative from the start, reflecting our commitment to establish initial guardrails while continuing to listen, learn, and work with the stakeholders most affected by these decisions.

The MARCH guidance laid an important foundation for our protections.

This guidance requires AI tools that use student data to go through New York City public schools' privacy and security review process, prohibits student data from being used to train AI models, requires human review of AI outputs, and makes clear that AI cannot replace student teaching, I'm sorry, student teaching, educator judgment, or qualified human decision making in high stakes decisions about students.

It also makes clear that any student use of AI requires educator guidance, critical evaluation of outputs, and age appropriate context.

During the 45-day period, which ended on May 8th, we received 6,491 survey responses from educators, families, students, and community members across our city.

We also gathered feedback through engagement with students, families, parent leaders, community and citywide education councils, as well as the Chancellor's Advisory Councils, the panel for education, and other stakeholders.

There were many key takeaways from the feedback.

Many respondents said they clearly or most or mostly understood the framework, and we heard recognition that AI and technology may have benefits when used thoughtfully and with appropriate safeguards.

Student in particular reflected both interest in AI literacy and concern about over-reliance.

Educators and staff also drew an important distinction between teacher-facing uses and student-facing uses.

But the feedback also made clear that our next set of policies must be stronger, practical, and developmentally specific.

The top concerns that were surfaced was the impact of AI on children's cognitive development, mental health and social emotional well-being, and data privacy and security.

Families and educators shared that they want stronger guardrails, clearer expectations, and greater transparency.

Families and educators have also raised questions about screen-based technology more broadly, how technology is being used and whether it is supporting active, meaningful engagement.

From this feedback, we understand that our current guidance does need to be strengthened to better protect students and respond to community input.

While we're still shaping our next set of policies, let me speak to the values of what is driving what comes next.

First, we need to center developmental appropriateness.

Across all grades, students need opportunities to read, write, problem solve, discuss ideas, and work through challenging tasks with teachers and peers.

We are mindful of cognitive offloading when technology does the work that students need to do for themselves.

AI should not short circuit those learning experiences or allow students to bypass the work that learning requires.

We also recognize that the developmental and instructional needs of students vary greatly from pre-K through high school, and our policies should reflect those distinctions.

As we build on the March guidance, we are looking carefully at how expectations should differ by age and grade level, how to prevent unstructured technology use, and how to prepare older students for a world in which AI is already present without allowing AI to replace their own thinking.

Second, we are focused on distinguishing between different kinds of tools and uses.

Not all technology use is the same.

The clearest risk is unstructured student-facing AI that allows students to bypass the thinking, writing, reasoning, and problem solving they need to do for themselves.

That is very different from a carefully reviewed tool designed for a specific instructional purpose used under educator supervision with clear limits and a clear role for the teacher.

It is also different from assistive technology or software like text-to-speech required by a student's IEP or 504 plan, a language access support for a multilingual learner, or a teacher-facing planning tool.

Those uses raise different legal, instructional and equity considerations, ones that we are factoring into our next set of policies.

Third, and probably most important, the safety of our students is essential.

Privacy, security, and instructional review must all work together as we evaluate technological tools.

New York City Public Schools has an existing privacy and security review process through the enterprise request management application or IRMA.

Software that has not been reviewed through ERMA is not approved for use with student or staff data.

In December 2024, New York City Public Schools added AI-specific standards requiring vendors to disclose AI capabilities, prohibiting the use of student data to train AI models and setting transparency requirements.

This review is essential, but families and educators are also asking broader questions about instructional purpose, accessibility, bias, and how tools affect the student experience.

As we enter the next phase of our work, we're working to establish a more comprehensive framework, one that responds to ongoing evolution of this rapidly changing technology and factors in students' protections, instructional quality, and more.

Finally, community voice must be part of the New York City Public Schools technology governance structure.

Families want more transparency, educators need clearer guidance, students need a voice in decisions about tools shaping their learning.

We're committed to engaging with each of these groups and others to ensure our approach to AI and technology reflects their continued input.

In summary, our March guidance established the foundation.

The feedback we've received since then from our surveys, from parent groups, and from our students has made clear that we need to build more detail and specificity on that foundation.

Any technology use must be developmentally appropriate, carefully reviewed, transparent to families, and in service of learning, not a substitute for it.

We are finalizing a stronger approach for the school year ahead so that educators and school communities will have clarity before the new year begins and will continue engagement around our updated guidance.

I would like to now turn to the proposed legislation.

Pre-considered intro 2026-21-29, requires New York City public schools to report on algorithm.

Oh boy, I knew this word was gonna get me.

Algorithmic tools approved for use in schools.

We support the goal of proposed legislation and pursuant to local 35, New York City Public Schools annually reports information on algorithmic tools in use at schools.

We look forward to working with the council and New York City Office of Technology and Innovation to ensure that any new reporting requirements align with existing legislation and our reporting systems.

Preconsidered intro 2026-2130 relates to quarterly reporting on data breaches involving algorithmic tools in schools.

Protecting student and staff data is of the utmost importance and is required by law.

We take any data breach seriously and follow all required procedures in the event of an incident.

We support the overall goals of the proposed legislation.

With that said, New York City Public Schools already maintains policies regarding uh privacy audits of third-party tools that process student data, including algorithmic tools.

These policies are publicly available on our websites.

The links are in the testimony.

And on the website, I would draw your attention to our schools at nyc.gov, forward slash AI, forward slash in our schools.

We also have data privacy and security uh compliance process as the place to search for them.

Moreover, New York City Public Schools reports all data breaches to the state in accordance with New York State Education Law 2D.

We also share information about recent data security incidents on our website, and that you would search that way with the forward slash at the end data security incidents.

We welcome continued dialogue with the council regarding the proposed legislation.

For intro number 412, this requires New York City public schools to disclose information regarding school admissions policies and procedures.

We support the underlying intent of the proposed legislation to ensure that school admissions processes are accessible, transparent, and well understood by our families and public.

As written, we believe the bill references information that is already publicly available, including each school's My Schools Directory page, which conveys priority structures, selection criteria, and prior year application and offers rates by priority group.

The same directory data is also published in spreadsheet form on the New York City Public Schools Info Hub for multiple years.

Comprehensive admissions process guides on the New York City Public Schools website explaining evaluation methods, including auditions and other assessments, and reporting required under local law 72, provide uh providing detailed admissions outcomes for each school.

We welcome the opportunity to work with the council to explore how this existing information may further be clarified and made more accessible for our families.

Thank you, and I welcome the conversation ahead.

Thank you.

I want to start with the student-facing AI.

I think it's one of the things that troubles me as a parent and a teacher the most.

We have some clearer screen time policies in place uh in our early childhood uh grade, you know, 2K through pre-K now.

But I'm gonna invite uh Dr.

Pate up to speak a little bit more to the specific AI.

And while she's coming up, I'd like to acknowledge that council member we've been joined by council members Maloney and Joseph.

As our first deputy chancellor said, we don't currently have a policy, but we are learning from insight across the nation.

We do know that we are paying very close attention to what it means to engage in developmentally appropriate use of AI technology, inclusive of screen time.

Thank you, Dr.

Pate.

You have approved Oh, Dr.

Hall.

Sorry.

Just to add specifically in the early childhoods, we just want to include that we have it's 30 minutes per week and 15 minutes per day.

However, it's not used for students to be on it alone, isn't that the intent?

The intent of it is more for a policy with the exceptions of like assistive technology for our students with IEPs.

And what age is that for?

And that would be for any early childhood.

So, two K three K4.

2K, 3K, yes.

4K.

So there's no policy for kindergarten through fifth grade.

No, that's the consistent policy.

Okay, there's no policy childhood.

Okay.

Um, there are seven tools that have, according to local law 35, um, that you utilize AI, although it seems that there are many more according to the vendor website.

Um, so I'm gonna ask about that later, but the simple fact is that you have approved the use of AI tools, or you have approved vendors with AI tools and schools to purchase whatever tools there are, and of course, you know, Gemini, the large language models are come preinstalled on the uh devices that we give our kids Gemini, Claude, and Chat GPT.

Um what evidence has the DOE used, peer reviewed journals, empirical evidence to demonstrate that these AI tools that you've approved as vendors are beneficial to our students and their cognitive development.

So when we're talking about um chat GPT, Claude and Gemini, we we have not opened these up for our student for student facing um only under some very um specific circumstances.

Uh Alan, would you like to jump in and perhaps Scott?

We also have Scott.

So I want this is a great opportunity to clarify that.

And just make sure you state your name for the record before you speak.

Scott Strickland, I'm the deputy CIO in the division of instructional information technology at the New York City Public Schools.

In relation to technology that's been approved for schools, we do have uh Microsoft products and Google products that have been longstanding approved as co-pilot and Gemini came out.

We evaluated them under our new standard and also approved them again for staff use.

I want to be very clear on that.

For students, there is only the Gemini system within Google that we open up to high schools if a principal requests that.

And there's about 35 high schools that have requested that this spring.

There are a number of tools.

I mean, I'm looking at this one right here, which says there are a number of tools on your web.

Canva is one of them.

There's another one, observe AI, a number, must be hundreds of vendors that use these AI tools.

And when you say open up to students, are you telling me that if you give a fifth grader a Chromebook, they don't go on to the school's subscription for Canva and use those AI tools?

I'll be clear on the on the devices.

So a Chromebook is a an user device.

Correct.

It comes with the Chrome browser.

Um when the student logs into the DUE network, we filter a lot of websites so that they do not have access to them.

In those cases, that's how we can control access to something like ChatGPT.

I don't know if Canva is open to students, that's really just a presentation uh piece of software.

Our teachers use that extensively.

Your student there are a number of vendors uh on your website that are approved vendors that have AI tools that are of available for schools to buy.

But I just want to make sure I understand your testimony that schools buy these products, these subscriptions with AI tools, but students don't or are unable to use these specific components.

That is correct.

Have okay.

I have a number of teachers and uh principals who would disagree who would disagree with you.

Um and I would also add, I'm not sure if you're aware of VPNs.

Have you heard of this magic tool?

Absolutely.

Are you aware that they are used in schools?

In some cases they may be, yes, but for the the students are supposed to log into the DUE network and use the DUE network for their work.

Right.

Okay.

I think the challenge we're facing already is that you've approved vendors that have AI tools, student facing AI tools, and we've had uh reports from teachers that these are being used in the classroom, students use them where students use Chat GPT or Gemini in the classroom, and they're somehow getting around your filters.

But let's let's pretend for a second that um no student is able to access these student-facing AI tools.

You have still approved vendors that have AI tools, including according to your testimony, Google, Google Gemini, correct?

Correct.

And what empirical evidence or peer reviewed journals has the DOE looked at to determine that this is for the betterment of our students' education and their cognitive development?

I'll turn it to the pedagogic.

Okay, thank you.

Make sure to pull that up.

Thank you.

See here, red means go.

So make sure the red lights on.

I appreciate that.

Um good afternoon, counsel.

Uh my name is Dr.

Alan Chang.

I'm the supervising superintendent for high schools in District 79.

Um, where I want to start is um, as both um uh council chairs Dinowitz and DeLarosa have said, um uh we deeply deeply are committed to our public schools as places of learning.

And we know that oftentimes in public discourse, uses of AI tools, particularly unstructured large language models like Claude and Chat GPT have a way of shortcutting and short-circuiting that thinking.

And by design, we have limited and prohibited that.

Your question is around the research that is behind this work.

Um, well, my my question is that you are approving these tools, and I do want to read one to you from local law 35.

Okay.

What it says and how it conflicts with the testimony was just given.

And the question is why we're approving tools when there doesn't seem to be any empirical evidence to demonstrate their efficacy uh in student learning.

So this is open gen AI and teaching assistant tool.

And it says the tool purpose, the tool is used to generate responses to prompts entered by a student or a teacher.

Requesting the generative AI tool to compose text response to a text input.

So I just wanted to read that to you because that was That is correct.

Uh, that Gen AI tool has been deployed to a very limited number of schools, and any student that interacts with that tool has gotten parental consent.

Uh mostly in district 28 is where that's been deployed.

I'm sorry, we have to, I'm sorry.

For everyone in the audience, we have to listen, and if you have someone to say, you can.

What do we do?

What's this little?

This means yes, and this means no.

Um, but but the verbal um assents or dissents uh cannot continue, okay?

Um so you in just one school district, you're using this tool, district twenty-eight.

That's the one where it's been piloted and is most heavily used by a small number of teachers and a small number of students.

And with what guardrails or or or uh assurances in place, this isn't uh part of cognitive offloading that so many of us are concerned about.

I'll make some general statements around that.

It's used with the But I I what I'm I mean, my general question is.

I just want to make sure you understand my question is that our kit it seems like even if it were just this one school that we're using this generative AI text.

Our children are being guinea pigs, and it just seems to me that they're I'm not hearing like any studies or data that the DOE has done or that a third party has done it or that peer-reviewed journals to demonstrate that these are tools that are beneficial to our students.

It seems to me that vendors have been approved and vendors are telling you that these are beneficial tools, and I'm still not getting an answer to my question, which is what evidence is there that these tools are beneficial to our students.

So, Chair, I'd love to be able to respond to that question.

Please, thank you.

Because um uh in our review of the evidence, um uh with our multiple meta-analysis randomized control trials in peer-reviewed scientific articles that state that um uh when these tools are being used with educators that are intelligent tutoring systems that are structured systems that are able to be in place, what they're able to offer is really an ability to be able to provide careful supports around instructional design feedback and adaptability.

We want to be able to cite one such article that is in the journal PNJ, Science of Learning, that had 28 studies, that was published in 2025.

There's another research article in PNAS that is a peer-reviewed randomized control trial.

Each of these pull together many, many hundreds of studies and analyze this research.

And this research ultimately tells us one thing that it is really important that students are not having unfiltered, uncontrolled, and unsupervised AI use.

But when it is being used together with teachers under guidance and supported, that challenges and asks students to be able to show their thinking and learning, that learning is helpful and does persist.

Well, that's any educational tool, show your learning, right?

And any educational tool is teacher guidance.

That's the purpose of a teacher, right?

Otherwise, what the hell do I need a master's degree for in teaching, right?

Um is it your understanding that that's actually happening in schools that teachers are guiding the use of AI?

And for what purpose is this AI used?

What's the educational benefit?

Well, I think the there's so many different versions of how it might be used.

The most important of which is exactly uh what Dr.

Chang just talked about with adult oversight.

So we know we have some tutoring uh programs that um and online programs that are adaptive, um, whether it that we're using more in supplemental and intervention spaces.

Uh I also want to go back to the conversation about like the chat GPT.

You know, we have these policies in our schools, and as and and actually is how come we felt it very prudent to put out this guidance with guardrails to be very clear about the IRMA process, what stood what it should and should not be student facing, what can and shouldn't be teacher-facing, and now as you know, I've spent the vast majority of my career focused on implementation.

And so that's where we are now.

We are how is this landing in schools?

How is the guidance and our policies actually coming to life?

So when we hear these examples about the misuse against our policies, like we want to engage with our superintendents and principals to understand like where's the misunderstanding in the policy?

We should not be having, you know, this particular, you know, the chat GPT example freely for children to be using in classrooms.

Um we should make sure that the Google Gemini is, we understand it to be the 35 schools, and if it's far exceeding that, those are great examples that should be escalated to us for further um explanation, and that's part of our implementation process, the chat balances of it.

I I think we hear all the time about escalating um issues, and I will tell you when it's almost 1600 schools escalating these issues to me.

I've had multiple principals tell me, you know, when it comes to this AI stuff, they're f they're frightened and they feel like there isn't enough guidance and they feel like they're building the plane.

It was an as they're flying it.

That was an actual phrase, multiple principals have said that they feel like they're flying the plane.

I want to share a reality with you that I experienced uh in high school about credit recovery.

You know, all of the principles are really incentivized to uh make sure kids graduate at any expense.

And so what they would often do is what's called credit recovery.

Does this sound familiar?

Credit recovery.

They'll plop a kid in front of a computer, I think in one of my schools, it was called Atlas, and they would sit for 45 minutes doing like online test stuff, and that was bad enough.

No teacher interaction with 30 kids in a credit recovery class.

And now I add the the we didn't even get to the biases, the hallucinations of generative AI, and we we do not have the guardrails right now to ensure that our yearning to make sure kids get these credits doesn't conflict with what good education looks like.

And this is, and it's frightening to so many of the students and families, and honestly, a lot of principals who just don't feel like they have the guidance yet.

Yes, these yet these tools, these approved vendors are still permeating their schools.

I'll ask two more questions and I'll turn it over to my co-chair, my neighbor from the south.

Chair Diller.

For those of you that don't know, my district is like here, and her district is right to the south of mine.

A little geography lesson for all you.

What is the DOE's position on AI detection tools given documented false positive rates that disproportionately flag multilingual learners and students with disabilities?

Can you repeat your question?

Yeah.

Does the DOE have what is the DOE's position on the AI detection tools given documented false positive rates that disproportionately flag multilingual learners and students with disabilities?

Or does the DOE have an approved AI detection tool or policy regarding AI detection tools?

AI detection tools, Scott.

Thank you.

You just stay.

Just stay, Scott.

We should pull up 13 more chairs.

Fair enough.

We do not have any AI detection tools that I would be aware of.

That are detecting if a child uses AI to to write questions, answer questions, write papers.

I think I'm getting tripped up on your question around students with disabilities and multilingual learners.

The detection tools creating false positives, saying particularly students with disabilities or multilingual learners are being flagged uh disproportionately as using AI with false positives, meaning accusing a student with an IEP or multilingual learner of using AI when it has not been detected properly.

But it sounds like the DOE does not, according to your testimony, doesn't utilize AI detection software or have any AI detection software vendors approved.

Correct.

But you do have plagiarism detection tools.

Those have been around for many years where students upload their papers and it detects for plagiarism, correct?

Yes, I am aware of plagiarism detection tools.

Okay.

So you don't have any such detectors approved for use in disciplinary or academic integrity hearing, uh, integrity hearings.

I mean, you know, even before AI, we we have very, you know, high standards for academic integrity.

Um we are not using a detection tool, but we want to make sure that our teachers are very closely connected to their students and they understand that this is their product or not their product, and how to make sure that the students are, you know, responding in a way that is theirs and not but uh into in other words I mean it sounds it's not I'm sorry to interrupt it.

Just sounds to me like you're talking as if it's 2015.

You're you're talking in terms of like, you know, good teaching, and but we are in a completely different world with the advent of AI, and tomorrow it's going to be completely different, and in two months, it's gonna be c it's gonna be astoundingly different.

And so I just want to ask one more question.

I'll turn it over to my co-chair.

Uh, you spoke about you know teacher intervention teachers, you know, uh looking at the work, knowing their students.

What training does the DOE provide?

What certification does the DOE provide?

Um, what how are you actually measuring whether teachers are trained in in any use of these AI tools that they are then being asked to use with students or in in detection of any AI generated content that their students are turning in?

For training, I will invite uh Dr.

Pete and uh or Tara to join in.

But what I would say is some of the one some of the programs that we rely on most are connected to our tier one curriculum or supplements to it.

And so the training for all aspects of the program is part of like the New York City Reads and New York City Solves Professional learning Plan.

So there's centrally uh designed professional learning, there is district level, and then there's school based.

So if a if a principal or superintendent or centrally were recommending a program that would you know would be good as a supplement for or an intervention program, it's going to come with training, as does all the all the aspects of that program.

So for example, you know, one of one of the programs that we know and we've gotten you know a lot of questions about and feedback is within HMH, there's writable.

We also know lots of our schools adopted a mirror for some supports and in literacy and ZERN and math.

Just some examples.

And so nobody would be asking teachers to engage in those programs without training.

And but that's also our non-computer based assessment.

So when it's part of the instructional strategy and either tier one intervention or supplemental, that is all part of the professional learning package that um is cupping into schools.

HMH includes required training for use of AI in the classroom?

No, it doesn't, it's not required, but there are some aspects of it that um some aspects of the program that uh different schools or districts are leading into.

But you're you're asking individual schools to choose and figure out, you know, whether or not they they should provide this PD to teachers.

No, but it I mean, if any program that a principal or superintendent or us centrally brings in would come along with professional learning, related to generative AI.

It depends on the pro if if that is part of the instructional strategy.

What is the question behind your question?

I'm just going to do your testimony seems to be a lot about like what you know what teachers are doing in the classroom, what schools are doing, and it just I'm just a little confused.

Like AI is not in the class, student-facing AI is not in the classroom, but also uh HMH has student facing AI and they get the training uh for these AI tools.

I just I'm just hearing a couple different things, and I'm just trying to get clarity on it.

FDC, if I could chime in just for a moment.

FTC, that's a good that's a great.

If I had initials like that, it's great.

Um, thank you for the question.

We want to recognize very publicly that AI is new, and in the largest school system in the country, we are taking it all in and we are navigating this evolving landscape.

To answer your question more targeted and more specifically, we have created a central-based AI guidance training for staff, and to date we have more than 8,300 staff members who have engaged and completed this course.

One of the positive things about this particular course on our system we learn is that as new information develops and evolves, we get to go in and to update that system and that training, and it actually pings educators to tap into the course again, that a new um uh a new module has been developed to further educate them.

I would acknowledge that there are more opportunities that need to be taken advantage of, which is why we are partnering extremely closely with our labor partners, UFT, who are doing cutting-edge um teacher facing professional development that is deeply building their capacity around appropriate pedagogical uses of AI in the classroom.

So, from my standpoint as chief academic officer, it is that we are engaging in a wraparound, rapidly evolving capacity building approach to what it means to educate teachers around appropriate use of AI in the classroom, and also we are starting to deeply look at what AI literacy should look like, be like, for students as they navigate what it means to be career ready, um, technology-facing, and also utilize uh AI technology um day to day in the work that they do from teachers.

Is it perfect?

It is not, but are we aggressively acknowledging that we need to continue to develop and to do that work?

We are.

The tools, I mean, you said we're taking it all in now that it's rapidly evolving.

That's true, but while we're taking all in, we're putting it out into the classroom without all of the trainings in place and without this digital literacy or AI literacy curriculum with any sort of evaluative tools to know whether our students are AI literate before throwing the tools at him at at them.

Uh I'm gonna I'm gonna turn it over to my co-chair, council uh Chair Delarosa, and also acknowledge that we've been joined by council member Farias as well as Aviles on Zoom.

Thank you, Chair.

Um, thank you for the line of questioning.

I think that you know what your dinos was just mentioning and the responses is why we see that the parent community has come together uh across the city and in CECs to ask for a moratorium.

Not necessarily because, you know, uh there because there are concerns, but also because there are huge gaps in our understanding of how the technology is being deployed and when, and um, I think we do need the time to be able to wrap our arms around what is happening in our classrooms.

I wanted to ask specifically about you uh said that in your testimony um that 6,491 folks uh submitted feedback to the AI guidance, correct?

Yes.

Um are any of those comments made publicly available for us to see?

They are not currently made publicly available, but we will get them.

It is our intention to share that to make them publicly.

Yes, great.

Does the DOE monitor student computer use in classrooms through programs like Go Guardian or any other monitoring tools?

For additional apps, you take that?

Yeah.

We uh some schools can purchase and use GoGuardian, they use the feature that monitors that the students are on task with whatever uh ed tech program is happening in the in the classroom.

There is a and just to be clear there's a separate module of Go Guardian uh called Beacon that is not approved for use and is not used in any of our schools.

And the devices, if the devices are take-home devices, it's monitoring them while they're at home too.

No, only in schools.

Only in schools.

And who has access to that information and what information is available through Go Guardian?

The purpose of it is to monitor the real-time status of the device as they're in the classroom for that session.

When it's turned off, all that information is not stored anywhere and used for any other purpose.

So can you just walk me through it?

So my kid is sitting in class, she has her computer open, she goes to YouTube.

What what do they what do they see?

She's not supposed to be on YouTube, let's just say.

Okay, and so that they can uh target.

They can walk by follow-up and say, please come back to your uh assigned tasks.

Are there any other monitoring tools like Go Guardian that are being deployed in our schools?

Any other tools like Go Guardian?

That's the one that's most prominently used.

I'm not aware of any others that you are used prompt uh across the system in a large way.

And again, they're purchased school by school at this point.

Okay, I do think that like you know that response of that I'm aware of causes concern.

We we want to know, and we need to know if there are other monitoring tools.

So it would be helpful if we could get back to the committee if there are others, it would be helpful in us understanding.

Um so parents have asked about approved AI products that are currently being used in the schools.

Uh it is my understanding that the DOE has declined to provide some of that information, and that some parents and we're asking, have parents uh do done FOIO requests uh for that information, and why is there hesitancy around sort of providing comprehensive lists?

Thank you for that question.

Um I'm we are currently working as part of our next space.

We're currently working with our superintendents and principals to have more of exactly what you just said, which is we need to understand everything that's in our schools, how they're being used, especially what the top ones are, which ones might be leading to some impact that's interesting because it is also new.

Um, and we do want to encourage schools to innovate and then scale, you know, what's working.

Um with that said, uh, we're in currently engaging in that process.

But I would say, as a former principal, that um any tool that I'm using, any curriculum I'm using in my school, a parent has a right to ask and deserves to be answered.

So, in cases where parents might have a question about what you know, my child uses X, you know, I didn't mean Twitter X, like insert any example.

Um, you know, can you talk to me a little bit about how it's used, how long they use it, you know, are they using it alongside a teacher?

Those are amazing questions for our parents to be asking.

We encourage them to do so, and we expect our teachers, our leaders, our superintendents, and ourselves to be able to answer those questions, especially if we've adopted them so locally.

Okay.

That's good to hear.

Um, does the DOE um register all contracts through checkbook?

The I will have to get back to you to make sure I believe the answer is yes, but let me find out and come back to that question either in this hearing or shortly after.

Great.

Does the DOE have non-disclosure agreements with vendors?

If so, why?

And are those contracts also registered with checkbook or passport?

Thank you so much.

Um at this point, I'm gonna turn us over to our chief privacy officer, Dennis Doyle, to jump in.

Thank you.

Yeah, thank you for the question.

Um I'm not sure exactly what you mean by non-disclosure agreements.

I mean, we have data processing agreements with all vendors that process student data.

Um, if that's what you're referring to, then the answer is yes.

Um if there's something else.

Um, I mean, I think non-disclosure agreements are pretty standard.

So is it a non-disclosure agreement or not?

Well, it depends what we're talking about.

So describe what you're describing.

So for a data processing agreement, which we have with all of our third-party vendors that process student data, there's non-disclosure aspects to that agreement.

They have to agree to keep our data, our information confidential.

So that is in some in some ways, it is a non-disclosure agreement.

We do not have any non-disclosure agreements with vendors that would prevent us from disclosing the existence of a contract with that vendor, for example, if that's what you were referring to.

Okay, or like what um you're procuring the technology for.

Correct, yeah, we would not have NDAs for something like that.

Okay, and again, we'll the question around checkbook and passport, we'll we'll get back to.

I can answer.

Okay.

Yes.

All of our um all of them are registered in checkbook and with the controller.

Okay.

Thank you.

Um the AI guidance states that on page six that a playbook is planned for June 2026.

What is that playbook and is it currently available?

Thank you.

Um, as part of our original plan, uh, we were releasing the guidance going through 45-day feedback, um, and it would result in the next step of a playbook.

That is how uh we started this process.

Given feedback, given some additional educational and professional reviews, and given the shifting national conversation, which has really escalated over just the last couple of weeks alone.

Uh, we are not necessarily committed to making sure that playbook comes out in June, because what we are committed to and was part of our testimony is going back to what that like baseline guidance is and flushing that out to be more responsive to the feedback, um, to take in some of the shifts in national conversation and research that's coming out.

Um, and so right now the playbook as it was originally sort of part of the original plan, is shifting.

Um, I I could, you know, invite Dr.

Pate if you want to add to that, or if that's just the answer, we can move on.

FTC, I think you did a wonderful job at responding to that.

What I would add is that this is not about a rush to get a playbook out for the sake of playbook.

What we are committed to is galvanizing the voices of community, which FDC talked about connected to the 6500 plus feedback that we received, and some of that feedback really shared with us a bit of a slowdown in order to really be thoughtful and strategic about what it looks like to ensure privacy, safety, and to ensure that what is instructionally sound for students is incorporated into what our playbook would be.

So the delay is not about getting the job done, it really is about being responsive to what the feedback has told us over this past 45 days, and to just acknowledge that this is really an iterative process that you can expect lots of loops, lots of changes, lots of shifts as we become better informed as as research continues to evolve.

And as we have these hearings, we've also give us like some really good insight about the direction that we need to go from some of your constituents as well.

Great.

I'm gonna ask two more questions and then I'll pass it back so we can open.

But um, I wanted to ask, I do have more questions for for later.

I want to give other people a chance.

Um, do you all have a specific policy around YouTube?

Scott, would you like to talk about YouTube?

Yeah, the policy on YouTube is it is not available for use generally, it's blocked unless a particular school asks for it to be opened up for educational purposes at the principal level, and we do that when that's requested.

Okay, we have heard concerns from parents that um there is instruction that includes you know, go watch this YouTube video on X, Y, and Z.

So we want to have clarity around um those type of instructions.

Um, I don't know if you have any feedback on that, but we have heard that from parents.

I mean, I I am aware and have seen where uh YouTube or some video link, you know, is helping bring a lesson to life, tapping into children's like immediate engagement in that.

That's part of the work that we're thinking about even on screens and screen time.

Um that might lead there.

But I have not seen an example of uh just you know, independently go watch a YouTube video.

That would be under those cases where we're saying that's the kind of work that is in either a whole class or small group, and certainly under um the direction of a teacher in the space.

Okay.

Um the other thing I wanted to ask.

We heard from some of the parents about this app, and I'm forgetting the name, that basically tell students that if they pay, like if you enter a credit card and you pay, you gain some powers.

You know the name, I don't know the name.

Play play prodigy.

Um, can you discuss if that is uh approved uh application or program that the DOE has approved?

Thank you for that question.

Um I believe play Prodigy is not an approved app.

Well, it actually is.

Oh, it is it is the free version is approved.

Okay, it's been through our uh process.

Uh there are other applications that have been approved that have sometimes have that feature uh of pay for more things that's not approved, nor is it encouraged.

This particular app is only used in a few very few schools.

Okay.

Um it would be interesting to figure out which schools and for us to see that information.

Um, obviously there's concerns there given the breaches and everything else that has occurred, um, you know, putting credit card information and all of that stuff at risk, but also um uh making sure that that is not something that we're encouraging our children to do.

Can I just clarify one thing and correct me if I misunderstood you?

So there is uh the version that's free, but we would never require our families to move into the pay.

That would be a choice that they made, and it we would I would not want to see our schools requiring that or encouraging that.

Okay.

More more on that.

I'm gonna pass it back and then we'll come back for round two.

Thank you.

Thank you, Chair.

Thank you.

I I I would just add your 6,000 uh similar responses is like less than one percent.

The parents and you have uh that's not a question, I'm just I'm just sharing that that's that's a I would say a very low percentage.

I'm sure many council members will have uh a lot to say about the people on the uh who helped develop the guidance.

Um and that's compared to the parent survey, which has significantly higher, and you do a lot of work, a lot of good work to make sure parents fill out that survey.

Um, so I would think a response sort of less than one percent isn't ideal.

Parents have a lot to say, rightfully so.

This is the future of our children, the future of our city.

And I I couldn't agree with you more.

I think you were given the amount of conversation that we're having about it, we thought I certainly would have thought we had more, but we did intentionally then make sure that part of our feedback processes included specific conversations around this with advocates with labor, with uh parent leadership structures, with our, you know, our elected.

Um, and so we, you know, and and really making sure we're listening, even the pet comments uh became feedback for us.

Thank you.

I wanna uh recognize Councilmember Zhuang for questions first.

Thank you, Chair.

Actually, I'm very disappointed.

They um early child education did not show up.

No, they're here.

They're here, yes ma'am.

I'm oh okay.

Very good.

Um the on the statement said early child uh division don't have any purview of gift and talent program, right?

But uh how come you guys have preview here about uh early child education program, which also run by DOE?

You guys showed up or mayor's announcement about pre-K 3K, which are also DOE programs, and then now you uh choosing uh come to gifted and talent program you guys don't have any purview over?

Is that terry picking or I I don't know how who gonna answer my question?

Veronica Coyer, um, I believe you're referring to the mayor's office of early childhood education.

Yeah, and the letter.

The division of early childhood education is involved with gifted and talented, however.

So, no one come from administration said to testify about my bill.

We have members from uh the division of early childhood from New York City Public Schools with us today.

Okay, they're here today.

Okay, I have questions.

So who can answer the questions?

Um, myself and my colleague Liana will support with that.

Okay.

Only 12.5% of children with disability in New York City schools are white.

What percentage of tuition payment for students with disability at oh this is not for you guys, right?

Um, uh what effects what effect does participation in elementary GNT program have on getting into specialized high school?

Do you guys have information?

We don't have that information at this moment.

Okay, but I did the research.

I have the information.

They have 12% more chance kids getting to uh specialized high school, and also um that program really benefited low-income students in New York City.

And next question: what effect does being a gifted and talented program have on those children later in their life?

We don't have that information at this moment.

Okay.

Have been in gifted and talent program increase it expected wage by 9%.

Do you know?

Do you have information about what current gift and talented program being black and the brown community?

Do you guys have any data or you can help me to get some data?

Council members Wang, I uh do you have data for us to share?

I I don't want to.

I was just gonna say we uh we did not bring the team uh prepared with data specific to gifted and talented today.

We brought the team to speak to AI and screen time uh more so we can accept your questions and note them and also follow up for the spaces that we're not specifically ready to answer you.

But it will take them back.

This is in the hearing's agenda, right?

About two bills.

I have to think.

But you guys did not prepare to answer my questions.

A lot of family uh depend on the answers.

They want to know the answers.

This is how this is the shows again, why those spatial needs family get ignored and the gifted and talented program get ignored, spatially in black and the brown community, and then now even doesn't have people come here to testify.

This is how we're gonna respect those families.

Um we do have data publicly available on our website that shows the demographics of students enrolled in gifted and talented programs.

And the by neighborhood too, by district, we have the number of students from every district who apply to gifted and talented and receive offers every year.

Can you tell me for the record how what's the ratio in Mahad and what's the ratio in Brooklyn?

I'm sorry, as you answer, please state your name for the record.

Yeah, my name is Liana Wright from the Office of Student Enrollment.

Um, I don't have all of the numbers in front of me, but we can share the links to where they're posted on the website.

If no one comes here to prepare, then why come?

Um I think some of the um bill that you were asking about also is from the mayor's office who provided some written testimony um in response to your bill as well.

The response is my understanding.

It's it said there's no purview over the gifted and talented.

Yes.

So the no purview, that's it.

Didn't say why, where, who has a purview.

If they don't have the purview, which we don't know which department can answer my question.

Well, we run the gifted and talented uh screening and admissions process.

It is uh housed within the division of early childhood, and then it is for our children entering uh kindergarten currently.

So uh we if we don't have the data to speak specifically to that's not necessarily publicly available, we will follow up and get that.

And then that's the reason I ask on my bill.

We need to have that data available.

And then everyone say it's not in their purview, no one in charge of this data.

I don't know which department gonna work on this issue.

And also for the disability students, spatial needs students.

No one come to testify.

Just said we agree on your purpose, but that's no information at all.

You guys support or not support who's gonna answer the question?

Council Murzhong, I'm I'm just gonna ask the DOE.

There's a lot of you here, all with tons of data, and I think before you leave today, I think you can testify to what the data and answer her questions.

I understand you don't have it in front of you now, but before you leave, you should have that and testify to that.

Thank you.

And then this will be the last, and then we'll move on to Councilmember Aviles.

Councilmember might uh Suzanne Sanchez, she's chief of special education, and I am here uh to answer any questions around students with disabilities that you may have.

Okay.

Well, we're good we're gonna go on to the next um the next member.

Um, but any of the questions that she had previously asked, uh, before you leave today, we'll testify to.

And I think you could find it between now and then.

Okay, thank you.

Councilmember Avilus on Zoom.

I will come back.

Yeah, thank you.

Thank you so much, Chairs.

Um, for this important hearing.

Um, I guess I'd like to ask the administration around issues of privacy.

Um I was certainly one of those council members, council members who is deeply concerned about this what we don't know, what's happening in our schools related to AI and student use, direct usage of AI.

So I'm deeply concerned about privacy.

As you all know, on May 4th, the state controller's office released a really scathing audit, finding that the New York City public schools maintains no written policies for data classification, risk assessment, or backup and recovery as required by the NIST data security framework specified by Ed Law 2D.

So given that that law was passed in 2014, why has the DOE failed to develop these policies?

This making obviously these data breaches, which are profoundly concerning, more likely and harder to respond to.

Thank you for that question.

And I just want to let you know we have our chief data privacy officer with us today who's going to respond.

Yeah, thank you.

Thank you for the question.

I just want to start out by saying obviously student privacy is of the utmost importance to New York City public schools.

Um we did work with the New York state controller on this audit over the last few years.

We submitted a formal response to that audit that sort of explains that we comply with all federal, state, and local privacy laws.

In some cases, we go beyond what the law requires, but that we're always working continuously to try to improve our data privacy practices.

And while we disagree with some of the findings in the New York State Controllers Audit Report, we did accept and agree the recommendations that were made in that report.

To some extent, we assert that we already comply with the recommendations in that report.

And one example is actually as you're referring to is um the requirement that we align with the NIST cybersecurity framework.

This is something that we uh we use, where as you noted, we're required by New York State Education Law Section 2D to align with the NIST Cybersecurity Framework.

And just a point on that: the NIST cybersecurity framework is not necessarily like a checklist of requirements for organizations to follow, it's a framework for assessing an organization's cybersecurity posture and to continuously evaluate it and to try to continuously improve it.

Um and so we have been implementing the NIST cybersecurity framework.

With that said, we recognize that there are improvements that we plan on making.

One is to update our data privacy and security policies page, which we will be doing in the coming months.

We will also be issuing a formal written data classification policy.

We do have a data classification policy.

It's documented in a few different places, uh, but we want to make sure that we have a very clear formal data classification policy for that.

We're also, as has been mentioned earlier in this hearing, working with the schools to make sure we have a better understanding of the tools that they're using to make sure that it's in compliance with their data privacy and security policies.

Um another thing noted in the audit report uh was the timeliness of notifications.

Um we're also improving on that front.

I think if you read the report itself, uh we have in 2024 and 2025, uh there were five, the state controller found that there were only five of 102 uh reported data breaches that were not made uh in accordance with the timeline requirements.

Um but with that said, we are striving to improve that.

Part of the difficulty with notification timeliness is making sure that uh we're having our staff be fully aware of what their obligations are and that they're reporting it up to us uh in a timely fashion.

We also can encounter difficulties with working with third-party vendors and trying to get the pertinent information from them in time in order to issue these notifications.

We also will have translation requirements when it comes to sending out notifications.

So there are a lot of um hurdles that when it comes to issuing these notifications in a timely way.

But with that said, you know, we're still working to improve that.

And also on the training point, um, you know, we've also improved in terms of staff completion rate for our training.

Um in 2023, 24, it was 117,000 staff members completed our data privacy and security training.

And 25-26, that number went up to 138,000 of approximately 150 or 160,000 uh staff.

And this is also an area where we actually exceed the law.

We require it for all of our employees, which we think is a good policy.

Unfortunately, that also creates separate challenges in sort of ensuring that everyone actually is completing the training that has access to data so that we're complying with the law.

So, in short, you know, we are working continuously to improve our data privacy and compliance policies and practices.

We've accepted the recommendations of the controller, and we look forward to you know issuing these new improvements in the coming months.

So just just um to circle back.

I mean, your updates on uh security protocols are gonna be coming out in a few months.

Is that is that what you're suggesting?

On our data privacy and security policies, yes.

So that's one of the requirements under New York State Education Law 2D is that we have that posted on the DOE website.

It is publicly posted there right now.

Uh we want to improve that.

We want to improve it to make it uh clearer, more accessible, so everyone has a better understanding of what our data privacy and compliance policies are.

And I also just want to add, too, along these same lines, um, we've also in the last year or just over a year ago, updated our chancellor's regulation, which is chancellor's regulation A20 on student privacy.

Uh that regulation had not been updated since 2009.

Um, so we're, you know, we worked very hard on that to try and bring our our own DOE regulation up to speed with the current landscape of privacy law.

We worked with a lot of uh parents, community leaders, advocates, uh school leaders, um, incorporated their feedback, and we we feel came up with a regulation that actually um strengthens our data privacy practices, but obviously this is a continuous process, and we're going to be doing more.

Thank you.

Thank you, Councilmember Obulus.

And if you'd like, I'll put you down for round two.

Uh, okay, I'll put you down for round two.

Council member Wong, followed by Gutierrez.

Okay, um, thank you, Chairs.

Um, I went to public school um, one to learn facts and also to learn how to think uh think critically.

And uh I sent my daughters to public schools, also expecting them to think critically, not to just type something into a AI engine and then take the results for granted.

And I've been playing with this uh consensus, perplexity, chat GPT, and uh um I was asking about myself, like what do you know about me?

You know, and then a lot of times they're wrong.

Yes, and I'm concerned because one uh it the technology is already here, kids are already using it, whether you see it or not.

And my other concern is that teachers are using it, whether they're using it to prepare the lessons or where they're using it to grade student papers.

Those are very concerning.

But I would like to start with one thing.

When I see students, everybody has AI, it's in between the ears.

This is actual intelligence, and my message is that they have to use it rather than relying on some AI engine telling them things and they take it for granted or they believe it's true, and I claim that the given technology, a lot of results you get from AI engines are force.

Now, with that, here are my questions.

Uh I I read from the rep website, the OE website.

There's an AI guidance, and then there were there were parents could submit feedback until May 8th.

And there's feedback somewhere.

And I would like to read it.

I want to learn about what parents are thinking.

Uh where is all this feedback?

Can I see it?

It is not currently public, but we are going to make it public.

Can you please do that?

Or send me the link.

You have my email?

Okay.

Um, next my next question is at the same AI guidance, page six, it says a playbook is planned for June 2026.

Or in one week, we're gonna be in July.

What is this playbook?

Where is it and is it available?

Yes, that goes back to um when asked earlier about the playbook is that the original intent um and timeline we we thought after we launched the initial guidance that by the end of the year we'd be able to, you know, really use the feedback to get us to this playbook.

But the feedback actually is taking us in new directions, asking us to think about um a developmental sequence, what can and can't be used, when it can't.

Also, the conversation around screen time started coming up.

We're trying to pay attention to all the feedback that we got through survey and our public engagements.

Um, and so right now we are not racing into a playbook.

We are carefully considering how to build out the original guidance that puts us on stronger footing with clarity, uh guardrails, and specificity.

Okay, thank you.

Um I have time some time left.

Okay.

Um, do you do you uh do you have a policy for telling teachers and what to do with AI when it comes to planning lessons or grading papers?

Because I I know I know like for even right now in HR departments, they screen the resumes and reject resumes if they don't see a certain term or they don't see a certain uh experience, and then applicants get rejected without without having anyone looking at the resume, just rely on some AI engine, right?

So what what is the what is your policy on uh or will you have a policy for teachers on how to use AI and planning lessons or grading papers?

Councilmember Wong, thank you for that question.

Um I want to make it very clear that um in the guidance I was released in March, um it really clearly states that um uh assessments and grading are prohibited uses of AI in our New York City public schools.

Okay, uh when you say is what about planning lessons uh what stops a teacher from telling AI generate a lesson about revolutionary war in Saratoga, yeah.

I hope they get great lessons.

Um I think what we have to do is to make sure that um uh ultimately the lessons are content aligned, they're factually accurate, and they are deeply engaging with the learning that we want students to be doing.

And so we do not recommend nor encourage teachers to be using AI tools to be generating those lessons.

We want to make sure that that educator input and guidance is culturally appropriate and accurate um for those learning purposes.

And to one of the things that you said earlier, we really truly want to make sure that the intelligence is something that we're developing with our students through deep interaction, educator guidance and support, and that is what we strongly believe in.

Okay, thank you.

Thank you, Councilman.

I'll put you down for round two.

Squeeze in one more question or go to round two.

Let's do it.

Thank you.

Thank you, Senator Chair.

Councilmember Hanif.

Thank you, Chairs, and thank you all for joining us.

I I do want to just at the top share that I'm I'm supportive of AI uh in our schools, and of course, urge intentionality uh before our schools and students administer these tools.

Um and I also want to add that I'm one of the signatories of the letter calling for a moratorium.

Um I believe that the way in which AI curriculum is built should be guided by students, and whatever is put in place should be piloted or observed, followed by an assessment and feedback, a feedback loop, so that we're not just saying we're not just dropping the curriculum, the plan, and then just moving moving on.

Um we can't avoid AI.

We need to get be on top of it.

Could you share if any biometric uh recognition technologies are currently being deployed in our schools for educational or security purposes?

Thank you for that question.

I'm gonna let Scott jump in on uh the different applications he just has that.

Yeah, I'll uh have Dennis back me up.

We do not have any biometric detection uh tools.

And that's that's uh uh prohibited by law.

And then can you uh um uh can you share how is the necessity of an AI tool going to be evaluated before introduction?

Oh, thank you.

So uh as these new tools are coming, the question is how do we you're asking how we're gonna evaluate them?

Um I'm gonna ask uh Tara or Mia to jump in on uh some of the work that's underway.

So both saying yes to specific tool or or company and then evaluating them year one, year two.

So what we are working to develop is a tool scorecard that will give us a more in-depth assessment of not only the educational value add, but also impact meaning outcomes uh for students.

Uh we are also currently um doing some beta testing.

Could you share just in in more kind of clearer language what the what the testing uh includes, how that is being conducted?

It's being conducted through professional learning experiences, where um some teachers and school leaders, 50 teachers actually, and um school leaders, are engaged in um programs that lend itself to AI empowered tools, and we are assessing whether or not these tools have instructional value add, and then collecting surveys, talking to teachers, figuring out the different creative ways that they are utilizing those programs so that we can start to better determine how we can more effectively assess um the value add of a tool that may be utilized within uh New York City public schools.

And these 50 teachers, uh, what's the representation of so they crawl they span um from 6th to 12th grade?

In addition to that, um they span our core content areas, so math, science, um, ELA as well.

And studies.

Will there be another set of teachers?

The goal is that we, as we develop this this scorecard, we want to enlist the partnership of our educators.

It is really about like how we trust them to inform us more about the value add of a particular tool.

It is not about taking the word of the tool creator to say, hey, this is great and go use it.

No, what are our educators saying?

How are they utilizing the tool?

What are what are their thoughts around the value add of the tool?

And what are the academic impact outcomes that that inform us that this is in fact a tool that we should invest more time and energy into as a whole, not leaning towards what the vendor is saying is good.

I would imagine they would all say that they're great and good.

So we can't we can't take their word um for law as it relates to.

Because I have a few more seconds, could you share for a similar scorecard is being assessed uh with students?

So I'm gonna jump in, Tara, um to do this work.

So Tara, sure.

Um I'll come back for second round, but ending here.

Thank you.

Hi, my name is Tara Carroza.

I'm the director of digital learning and innovation, former District 75 teacher.

Um currently, we are not uh evaluating tools that are direct student AI use.

We're starting with teachers.

And the reason for that is because we have the same concerns that you as an advocate and many others have shared with us, particularly from the 45-day feedback window and the listening sessions that we engaged in.

And so we're starting with teachers and leaders who are already using the tools for the purpose of supporting their productivity, for supporting their lesson differentiation scaffolding.

We already have high quality instructional materials and grade level expectations within all of our New York City Reads and Solves curriculum.

And so we're looking at those standards that we've set.

We've vetted those curriculums.

We know they're research-based and evidence-based, and we're taking a similar approach to vetting our tools beyond Irma.

Thank you so much.

Thank you, Councilmember Hunt.

Okay.

I would I would recommend our students seeing some of this, being able to work with a tool and and for us to see how they're responding, so that they're also a part of this process.

Agreed, and we we are creating the same evidence-based tool scorecard for all of our stakeholders.

We're in development of those, so I did wanna add that.

One for families as well, so we can empower families to go through the same process.

So we're all looking at the same evaluation criteria.

We're not saying, okay, the DOE is doing one thing and not sharing that out.

We really wanna be transparent in our next steps forward with all, and of course, involving students and empowering them with that same agency and critical evaluation and critical thinking as Councilmember Wong shared.

Thank you.

Thank you, Council Member Honey.

Love the questions around evaluations.

FDC, give her a highly effective and 3D, please.

Very niche joke.

Okay, uh Councilmember Gutierrez.

Thank you, Chair.

Um I have a few questions, and I was one of the signers on on the on the letter.

At first I, you know, wanted to go through the guidance thoughtfully and was able to meet with some parents.

I think some who are here to testify and really listen to their concerns.

I think the first thing about the guidance was that even like the the process seems backwards where you're you put out a guidance, you're collecting feedback, and now you're gonna use that feedback for a playbook.

It just feels like the process doesn't make sense.

If you wanted to actively use feedback, you would have called for feedback first and then put together a guidance.

A playbook, I I don't know if a playbook equates to the same as guidance.

I just I want to make sure that like if you I don't know, I feel like we're creating something that we don't necessarily need to do at this point.

So I just want to make sure that I raise my concern with the process.

It really feels backwards.

Um, I think in speaking with a lot of parents from my district, from my community, especially for those that are younger that are in elementary school, there's concern for the fact that they're kind of being brought to the table months after some of these tools are already being introduced.

Uh, the top concern being with these tools being introduced into the classroom.

I'm sure you'll hear from folks that are testifying that today, but bigger picture.

I think that we deserve a vision, like a digital literacy vision from the DOE.

I feel that we deserve to know what are the expected outcomes that you all as a governing body of all our public schools expect to have before you start throwing around a guidance or a playbook or mandating vendors.

And I will say from just the day-to-day school um feedback.

I think a lot of principals feel like they're being forced into interacting with these tools.

I think a lot of schools feel like they're not prepared and they're being told from the top from tweet that they have to utilize this.

So a lot of this has been very fragmented, has been fragmented from when you put the guidance out, it is being fragmented from the way that you're implementing it.

So I would just love for you all to share.

If there is a bigger picture.

What is the desired outcome?

Um, so that we can tell people here publicly because they are concerned?

So, yes, we want a moratorium, we want to slow it way down, but I think a lot of that is your fault.

I think a lot of that is because of how you have laid this out.

So, what can you tell me now that's gonna help us feel like we can trust and feel calm about the next four years under this administration and what you guys have laid out since March?

Thank you.

That's really important feedback.

And I think what I would want to mostly um offer here is I don't think we disagree.

You know, the prior administration, you know, hit the guess on go with AI and technology, and we were coming out of the gate.

This was very shortly after Chancellor Samuels came in.

You know, AI has been a moving train.

And so one of his first priorities, in fact, I think it was a week after he established his new cabinet.

Uh, we started on the fifth.

I think we put out uh the guidance on the 24th-ish.

Um, and that is because we recognize um that this guidance uh this this train was moving, and we wanted to put in guardrails and back to council member Wong's question, the some of the uh answers to your question are what is in red, what is in yellow, what is in green that answers some of the specifics.

We need to get more specific, so I want to acknowledge that.

And I think that's why we are also saying today and sharing with others, like we're not moving fast into this playbook.

We want, you know, Chancellor Samuels.

I'm sorry to interrupt what um if you can answer this in your continued answering is what is the role that the playbook is going to play that the guidance that you all put out did not where we hoped I think the original intent, and this was you know, uh back to the original plan, Dr.

Pate.

If you want to jump in on the original attempt of the playbook was to take this feedback and make it very much like answering um to like the questions that were coming in and feedback and realignment, but I think we need to go further than what the playbook's intent uh was going to be and kind of go back to the guidance and ensure that we're doing that engagement uh that you're talking about.

This is a core value of Chancellor Samuels, of our mayor, of this administration.

We believe that we do need to reset using the feedback and also the whole conversation around AI and um has shifted nationally since our guidance went out.

Um I'm not saying it was our guidance, but um it is important for us to take all of that in so that when we get to a vision for the role that technology, screens, and AI, because they're they're all intersecting but different, uh, that we're able to offer you that, and I can share with you now that that is the a current process that we are engaging in as a team.

Okay, can I just ask one last question?

Did you want to think it's a yes or no answer actually?

Yeah.

Is there a world where after this playbook is done, the the AI original guidance that you release in March is tossed, is garbage?

I don't think so.

I think we I we stand by our March.

No, that's okay.

Thank you.

Thanks, Chair.

Thank you, Councilmember Gutierrez.

I I while I can appreciate you're not moving fast into a playbook, you want to be thoughtful.

I think the evaluative tools and the feedback from parents, students, educators are important.

These tools are in the classroom.

Yep.

So it it it is uh uh it's astounding that both of those things are true at the same time that we've already released these tools into the classroom without without the playbook, without without the guidance for the students and the teachers that are really required um before engaging with these types of tools.

Uh I will uh now recognize council member Joseph, followed by Maloney.

Great, thank you, Chairs.

Um good afternoon.

Um, as we're having this important conversation around AI, we we've been in this space before.

Where are the parents in this conversation?

I heard nothing about parents.

I heard teachers, students, educators, but I have not heard where is parents in this conversation are they made aware of AI in the classroom?

Do they have the option to opt in, op out?

And another part of this conversation, I'm not hearing language access, and we know a lot of our parents in the city.

Lang English is not their first language.

And I have a question for the champ, Deputy chancellor.

Who are your two leads in this AI digital learning space right now?

Who are your two leads in your office?

Within the thank you for your questions.

You're welcome.

You want me to lead off with that one?

Is that Mia and Tara?

Within the division of the first deputy chancellor in academics and instruction, Dr.

Pate and Tara are leading that work, but I add like that work hinges upon the partnership with DIIT, the partnership with data privacy.

So these gentlemen also are very integral in this work.

These things cannot happen without each other.

That I know you have preaching to the converted.

So I've asked my question.

Okay.

So tell me, how are you taking in consideration when you deploy AI tools?

How are parents made aware?

How are you using parents' voices related?

Parent engagement, education that cannot cannot happen without parents.

Where are the parents in this conversation?

In addition to the AI guidance and the feedback loop, superintendents were provided with on the ground resources, and superintendents did their own engagement, unique to their communities, where they utilized the different governance structures within their district to engage families to collect insight, to collect their feedback, and then they provided us those trends, the themes, and that was incorporated in addition to the 6,400 researchers.

Which we still consider as a low number as having the highest, the largest school system in the country, 6400 is very little.

So here's my question.

When you took that feedback, what were the three top things that stood out for you?

So the top three concerns that stood out from the cumulative information that was gathered was around cognitive offloading, was around safety and around privacy.

And that's been our theme, privacy, cognitive and development.

It's a national theme, so yes.

Correct.

Um, so we're gonna put that there with the parents.

So you know when you do your playbook, even council members Gutierrez said you should throw it out.

But parents have to be center of this conversation.

And what about our educators?

What PD is being provided for them?

Because I also know that some of these curriculums also come in a hybrid model where you have a technology component to it.

How is that being used and how are teachers being trained?

So we spoke a little bit earlier about the two-tiered approach that we're taking now.

We have developed AI guidance at the central level, and our educators are engaged in that training.

So to date, more than um, more than 8,300 staff members have taken that um information and that course.

In addition to that, on the other side of it, is that we are partnering with our labor partners who are doing some really, really great work.

Even myself, I've attended uh their capacity building training.

Tara has attended the capacity building training that uh the UFT is offering.

Um so that is another way in which educators are being trained very skillfully around that work.

What's the timeline to make sure that every educator somehow went through at least one session of these classes that they're not navigating?

Because I know some kids I could tell you are running circles around their teacher with the use of AI, and they can find other detectors that can make sure that the the product that they produce is AI, you cannot detect if it was used through AI, and I know quite a few.

Yes, so I have uh 11th grader at home, I have one too.

Who is who is currently providing professional development to me, uh connected to um all that AI has to offer, and then some so I do agree with you around the need to continue to build capacity.

We have not set a timeline as of yet that I can offer you, but the technology is moving faster than you're working.

By the time you go to sleep, somebody has created a new one, and you are absolutely correct, and that there lies the challenge in it.

So I don't disagree, I am agreeing with you that because of the rapid evolution of technology, it sometimes feels like we are chasing the development of professional learning and capacity building opportunities.

But we are actively creating and developing that.

And and one thing we want to address with AI, how are you addressing biases that do exist in existing um AI tools now?

So currently there isn't a solid tool that we have endorsed that tracks or trace or keeps account of bias, but we do know that it is again a mega problem, that we are looking for resources that help us help us do that.

But we don't have a solid uh tool that we are currently utilizing to engage in in that work.

I think where we where we can, where we can really kind of dig our hills in is around AI literacy development and awareness.

I think people have to understand that it can be wrong.

It is not always right.

And it says that too.

If you depending on the prompt that you type in, it will say this answer may not be accurate, you should do further research on it.

Yes.

So the development really comes through awareness to your point.

So the more we get that message out to educators, we get that message out to school leaders, we get that message out to students.

By spreading that message, we start to build an awareness that you shouldn't put everything and trust it all as is.

You still must continue to critically think.

You must um engage with AI technology, not not as a consumer, but as a producer, meaning that you are asking it questions that will probe your thinking, not to deliver and to give you blanket answers that may or may not be true.

So it really is about building awareness around a rapidly evolving, as you said, every day.

Um it's a monster at this point.

Yes, it is.

And how do we support our English language learners and students with disabilities?

And I have the rep.

Um, you can answer that last question.

So I I do want to invite our friends from Dial and uh multilingual languages to just kind of add a value add piece to that.

Thank you.

Thank you, Dr.

Pete.

Thank you, Dr.

Page, Suzanne Sanchez, uh Chief of Special Education, and thank you, Councilmember Joseph for the question on students with disabilities always at the forefront.

We know uh what you're doing, and hope you're gonna visit PS6 seed this summer, but that's another conversation.

Um students with disabilities benefit significantly from the use of assistive technology, and that is something that New York City public schools is committed to ensuring those students have the right technology in the right cadence in the right dosage at the right time.

Um, Chair, you brought up earlier around research and efficacy studies and the evidence, the body of evidence on technology for students with disabilities is irrefutable.

Uh, and it's important to note that technology for students with disabilities, what uh typically refer to as assistive technology recommended on a student's IEP follows a really careful and thorough evaluation process by an expert in that field who then recommends the use of a specific AT tool, could be a device, could be a software, could be an app, um, all of which are go through the Irma vetting process.

Uh research is very clear that children with disabilities who utilize AT not only see growth and improvement in academic outcomes, but mental health, social, emotional as well as we remove barriers that to learning, uh students start to feel better about themselves, and that is certainly seen and felt in our students with disabilities.

Okay, but I just want to clarify for everyone listening the specific nature of what you're talking about AI, you're talking about an individualized education.

Well, I know what it stands for, but I wanted to have like a little gravitas and uh you know emotional cadence of the word individualized.

Get your gravitas.

Individualized education program.

Was the dramatic pause there?

Was that good?

It was great.

Because what you're talking about is something akin to providing extra time on the test, and what we're concerned with is providing tools that give the kids the answer on the test and the cognitive offloading.

So I just want to clarify that for anyone listening today that absolutely if there's a specific tool for a specific student to ensure that they have access to the same educational opportunities as everyone else, that is not the same thing as throwing AI tools into the classroom with no vetting or little vetting, relying on vendors and corporations to provide guidelines instead of listening to parents, students, and educators.

Is there not a DOE approved digital literacy curriculum?

I mean, we've had issues with digital literacy.

We've had for well over a decade social media, we've had bias in in the media, we've had when I was in high school in the 1900s, we had.

It makes me sound right.

Oh, I see some other people who are also old as hell, like I am.

Um that we had lessons in curriculum just like this because the internet was coming of age and anyone can put anything on.

So, how is it that we don't have a digital literacy curriculum?

How is it that the DOE isn't requiring this?

And yes, update it as necessary, but how is this not like part of the general curriculum in our schools?

Yeah.

Um so, Chair, I I think this particular point, the dramatic pause around individual.

Individualized, thank you.

Is extremely important, right?

At no point in time do we ever want for there to be any tool that shortcuts thinking that our students are doing.

We know that.

But we don't want there to be, I'm not asking what you want, I'm asking what the policies are in place because we want the same thing.

I don't think you don't.

I just want to be clear on that.

And I want to answer your question about um uh digital literacy, because you you are correct.

State law requires that we are thinking about media literacy.

We're teaching for accuracy, biases, and facts back in whether that's the internet or Wikipedia, as students are writing their research papers.

It is a part of the curriculum that teachers are really thoughtful about engaging students in details about how we write research papers, how we're defending our facts and our citing where our work is coming from.

And so New York State Department of Education is actively updating that guidance to specifically require information around use of artificial intelligence, and we are working with our higher education partners to learn together with them how we're addressing this particular issue about teaching, not only I would say artificial intelligence literacy, but I would also argue that there's a real civic readiness that we have to be prepared for, because as AI really infiltrates and works into our entire society, it is dramatically shaping how we are confronting facts, how we're engaging in discourse, and is extremely important that when we want our students that become leaders in our society, that they're working closely with each other, with researchers, with our teachers, to be able to learn about this information and to be able to use it appropriately that is enhancing what it is that they're learning inside the classroom.

I love everything you said uh about what our students need to learn, what our teachers need to do, except that you don't need to wait for state guidance to do it.

You are empowered as the Department of Education to determine what goes into our classrooms.

You're there there are limitations and requirements, yes, but you have programs all the time, you have civics for all week, and you spend just a week on it, just a week on on civics to highlight what's going on perhaps through the year, but you have no idea.

In no way are we saying that we are waiting for the state to do that.

I said that we were collaborating with state ed to make sure that that guidance is updated.

We have tremendous examples of what is happening across our schools.

You have examples, I know you do, and I know you do, and I've seen it, and I've seen and I've spoken to, and I want to thank you for help setting up those meetings about the discussing the seal of civic readiness.

There are incredible things going on in our schools, and it seems time and time again you sometimes provide support for that and you celebrate it, but there doesn't seem to be an expectation.

Uh, for that that all of these wonderful things you're talking about happen in our schools, and it does not seem to be a clear evaluative tool to ensure that if even if you do expect it in our schools, that our teachers and our students are engaging in this type of learning, the very type of learning, trust me.

Yeah, I know you know, and I know that we need in a society so now that's seems so like at each other's throats that we absolutely need civics education, absolutely need digital literacy.

But it is your job to ensure that this is happening in our school, not just providing opportunities as a supervising superintendent that works with our high schools, that is helping to ensure successful implementation of future ready NYC as one example where we are working with industry partners, with our teachers, with college professors to make sure that is a key component, and it is data that we are monitoring as part of our instructional work to make sure that students are prepared for this.

I'm gonna turn it over to council member Maloney for questions.

Thank you.

First, I want to reiterate the urgency of this conversation, because as has been said, AI is already in the classrooms, it's already in our lives, and that includes the parents, the teachers, the students, the principals, and we all agree on the need for AI guidelines and frameworks and age appropriate use of AI in schools and in life, while preserving critical thinking, accuracy, student privacy.

And I want to flag that we need to be moving a lot faster because while we are slowing down, AI development is not slowing down, workforce is shifting and people are gonna be using this in everyday life.

So your testimony notes that there's changes that need to be made to the guidance, and some of them include distinctions between teachers' teacher use cases and student use cases, more rigor around the age appropriateness for use of the tools, how you introduce AI, at what age.

Um, so what what changes are expected and when will we get that revised framework?

So those are all of the things that we are exploring and hope to be able to include in upcoming guidance.

We are working furiously through that.

I cannot commit to a date, but we will be offering clarity on those exact items.

Um a follow-up question to that is this is a very agile technology that is gonna require agile governance.

So, what is the system for continuously revisiting these guidelines and the protocols because this is rapidly changing, and we, you know, we can't have a line in the sand, it's gonna have to keep moving.

So, what's your thought on the process for that?

I mean, I couldn't agree more.

This the way it's moving, we have to bake into what our next evolution of our guidance is, this iterative process that includes feedback loops and and like coming back to it, you know, in some sort of time frame, and that those are part of the conversations we're having right now.

Great.

And we'll as that go ahead.

I want to be able to add that also, you know, you and and many others have spoken about here is that we're doing this together with our educators, with our parents, and really with our students.

Um, and that that is gonna be a really important part of this and ensuring that we have processes that are regularly helping them understand, but also give input and insight into this work so that we can update and continue to be able to revise this work.

Great.

Um, and I would expect that the next set of guidance will have details on and on how we'll continue to iterate.

Um, and on the point of AI civics, I completely agree with the points that were made earlier.

Um I'm gonna make assumptions, but most, if not all students are getting access to smartphones, their parents and teachers are getting access to these tools, and they're, you know, they're readily available outside of the classroom.

So, how are we teaching students to use AI responsibly, critically, ethically, not becoming dependent on it, uh, viewing AI as a tool, not a replacement, and critically evaluating that distinction to interrogate AI, not just operate AI.

And I know you you mentioned this before.

Yeah, and I think that that is truly a conversation, and some of the most important things we could be focused on in education today.

And uh, you know, in an age in which it is so readily available for us to be able to get access information, but we can't understand whether or not that is officially true, or to be able to participate in our democracy requires that we're educating and working with each other on that.

Um, we have a number of teachers that have been very thoughtful about thinking about what this it's not just literacy, but really working deeply on evaluating um uh the information that they are getting.

Um, and we look forward to sort of working with our labor partners, with our school principals, and really being able to sort of showcase and demonstrate what that can look like so that we can help support that happening across more and more of our schools.

I would echo the oh go ahead.

I wanted to jump in.

As Dr.

Dr.

Chain, Future Ready NYC is within him and uh future ready NYC and the Office of Student Pathways this year did create a high school instructional model module for students, uh which was 38 teachers delivered.

It's a 15 um lesson module uh covering AI concepts, ethics, societal impact, and connection to future ready pathways that's happening at the high school level, and then uh for this coming school year, our computer science uh education team, which really historically has done a lot of the capacity building and developmental work around computational thinking.

Um they helped co-create our our state standards on the computer science and digital fluency standards.

They will be um doing uh a K-8 um support with teachers as well as instructional modules that aren't just use AI, it's really looking at the conceptual concepts across age and developmental grade bands.

So they will be implementing that within this school year.

Great.

Um, just to close out, Chair, um, I in the next iteration, I would love to see those specifics about age developmental AI civics and how we're introducing and probing students as their experience with AI evolves too.

And similarly, workforce development where college graduates certainly and high school graduates are going to be expected to know how to interact with these tools.

So I'm in the camp of we need to be introducing students at an age appropriate level with a lot more details to what we're in the uh in the guidance, um, but uh we are gonna be building this plane as we fly it in into the future.

So it's it's uh a problem that we're far behind on and will continue to be because of just how fast this is evolving.

So thank you, Chair.

Thank you, Councilmember.

Thank you.

I have a few more questions.

I wanted to ask, so local law 35 uh speaks to vendor disclosure, and it says that all agencies need to report the use of AI tools.

Uh and the latest report identified seven tools used by New York City Public Schools.

I won't go through all seven, but there was one that says annual professional performance review measures of teaching leadership practice calculation, vendor not disclosed.

Is there a reason why this vendor is not being disclosed?

Do you have a sense of what?

Yeah, I'll start and maybe uh I'm sorry.

You guys can continue.

So that's the the Mosul process that's been in place a long time.

I think the date on that is started maybe 2008-2013.

It's not really AI.

We put that into the report because it's an algorithmic tool.

It does make observations and then there's scoring that happens in the background.

Uh so that's what the the reason it's in there.

It's an algorithmic tool uh that came way before General Trivi uh hit the marketplace.

Mosul is the measure of student learning, which is part of the uh teacher evaluation process and uh principal evaluation process.

Well, not mosuls in the teacher space.

Yeah.

But regardless, why can't that vendor be specifically disclosed?

Yeah, we'll get back to that.

It's it's possible it was developed internally, but we'll get back to if there was a third-party vendor or if it was internally.

That would be great.

I wanted to also ask, I know OTI is not here, but in your perspective, um, what role does the chief privacy officer and OTI play in the operational process and um does DOE utilize a master service agreement procurement process for NYC schools?

Thank you.

I'm gonna bring in our data privacy uh to speak on his role and his work with OTI.

Yeah, I guess speak a little bit.

Thank you for the question about our relationship with OTI.

Um we work closely with OTI when it comes to the ERMA process and the data privacy and compliance process.

Uh one of the parts of that process is to have third party vendor products undergo a cloud review conducted by the Office of Technology and Innovation.

And we will also work closely with OTI and the chief privacy officer's office at OTI in responding to any cybersecurity incidents.

We'll work closely together and sharing information about the incidents and make sure we're responding accordingly.

I'll defer to somebody else on the contracting aspect of that question.

Anybody else?

Master service agreement procurement process?

I'm gonna have to get back to you on that.

I don't want to misspeak on that.

Okay.

Um OTI has a citywide AI governance, the AI action plan and the generative AI use guidance.

Uh does this does OTI's overall citywide guidance uh bind the DOE or does the DOE believe uh itself to be outside of OTI's authority?

I don't believe we see ourselves as outside of OTI.

Would you have any more insights to share?

Yeah, I mean, I think we we generally comply with OTI's guidance when it comes to issues like cybersecurity and data privacy.

And if there's a conflict um with OTI's suggestions or or guidance, who governs?

I think it would probably depend on the subject matter area.

I mean, for example, um there's the city's identifying information law, which is administered and uh by OTI.

That's something that the DOE is actually not subject to because we're subject to separate laws under uh FERPA New York State education law to do, so I think it would it would depend on the subject.

It depends on subject, okay.

Thank you for clarifying.

Um, I would just add a bit too uh this we uh to the extent possible love to uh be consistent with what OTI wants, both technically, legally, etc.

Every once in a while, as Dennis mentioned, there's things that are over and above on the educational setting.

Also, there's the situation where this the DOE is really a agency of the state.

Yes.

Um and the mayoral control, it can be removed at any point, and therefore we need to be able to operate independently, but we do like to follow what they're doing.

We we work with their cyber command a lot on uh general security and then specific incidents we work cooperatively together.

Thank you.

I wanted to just ask a little bit about um the parents' rights um under the parents' bill of rights for data privacy and security.

Um, with New York City schools and with the New York State Department of Education.

Um, how many complaints do you all receive per year?

What is the typical outcome of those complaints?

How are they sort of followed along?

Yeah, thank you for the question.

Um I don't have the specific number in terms of complaints that we receive uh from parents.

Um if we were to receive a complaint from a parent about an unauthorized disclosure, we would investigate it and if we determine that personally identifiable information had been subject to an unauthorized disclosure, we would follow our normal protocols.

We're we notify the state of the disclosure, and we then make sure that any impacted student is going to receive a direct notification of that disclosure.

The DOE's data privacy and security policy states that parents have the right to require written consent before personal identifying information in your child's education record is disclosed.

The policy also states that in some cases the law allows disclosure without parental consent.

I wanted to ask if you could clarify how a parent a parent can exercise the right to require written consent um before their child's education record is disclosed, and is there a specific form that parents must submit to require this consent?

So we have uh we have forms on our website that um are designed to um for parents to provide consent for disclosure of education records.

They're not required to use those forms, as long as uh we have uh the consent in writing from the parent and it meets all the requirements under FERPA, and meaning that it identifies the information to be disclosed, who's it's going to be disclosed to the purpose of disclosure and it's signed in data, we'll accept that.

But we do have a ready-made form available for parents.

Um, and then remind the first part of your question was oh, for just requiring consent in general.

The general rule is that consent is required before any student information is disclosed.

There are, however, a number of exceptions to that general rule under FERPA.

For example, if you're disclosing student information to other school officials, it could be other employees within the school.

That does not require consent as long as that disclosure is being made to another school official who has legitimate educational interest in receiving that information.

That's just one of the many examples.

I mean, there's a health and safety emergency exact exception, for example.

If a student is has a medical emergency, needs to be taken to the hospital.

We don't need to get consent from the parent before we start disclosing that information to you know medical technicians who are going to be treating the student.

Okay.

Thank you.

Thank you, co-chair.

Um I want to talk about this uh state controller audit that was raised previously by council member Aviles.

You familiar with it?

Yes.

Okay.

Good.

That was that was an easy one.

Um you spoke about the training previously about the required training.

Part 121.

Uh requires annual privacy and screen, and security training for employees with access to personally identifiable information, correct?

That's correct.

Okay.

And the DOE requires employees to complete annual privacy trainings, correct?

Yes.

So the report, the controller found that in 2024, only 73% of employees completed that training.

Is that correct?

That's right.

The audit found the DOE did not review the completion list to verify that employees with access to personally identifying information had completed the training, correct?

That was the finding in the audit, yes.

That was so it is correct.

So the DUE could not prove that every employee with access to student personally identifying information completed the required annual training, true?

Yes.

Okay.

So that was from 2024.

Does the DOE know today whether employees that are not completing annual trainings have access to personally identifying information?

No, I mean the the difficulty with this, and this is an area where we actually, as I said before, we actually exceed what's required by Ed Law 2D.

The difficulty that we have with this is we do not have uh a list of all of our employees that identifies in that list whether they have access to personally identifiable information.

But as part of your policy, you require all employees to take it regardless, yes.

Thank you.

And what percent of employees have taken this training?

Uh our most recent numbers from the most recent school year, we had 138,000 employees complete the training.

Um, what percent is that?

Uh I would have to get back to you on the exact percent.

I think it's out of roughly 150 or 160,000 employees.

So that's not a hundred.

It's not a 100%.

But I mean, if your policy is all employees and you can't identify which employees have access to personal identifying information, I guess it makes sense to just train all employees.

But even and that was uh in a report issued two years ago.

So even with your own numbers, you not ever and your own requirements, not every employee gets this training, correct?

Every employee receives the training.

Whether they actually complete it is a different question.

I mean, we we strive to have 100% completion, and we're working towards finding ways to make sure that we get closer and closer to that number.

It's extremely difficult given like the number of employees we have.

We have employees who are going out on leave, employees who are coming onto the street.

So you have roughly 130,000 out of 160,000 completed, is that was that what you tested?

138,000, roughly.

Out of 160,000?

Ballpark, yes.

Okay.

So you have 22,000 teachers going on leave, not completing it?

Not necessarily teachers, just the DUE employees.

You have 22,000 teachers a year going on leave.

Well, there's some flux in that number, but again, not all of those employees like necessarily have access to personal identifiable information.

Right, but you don't know which employees have access.

So you made a smart decision to require all employees to complete this training, but it doesn't sound like a requirement, it sounds like a desire.

So what steps are being taken to ensure that the adults, the the employees that have access to this information, get the training that you require.

Yeah, it it it is a requirement.

You know, whether or not we we can actually uh make sure that everybody in 100% completion is the challenge.

We send out the annual training every fall at the beginning of every school year.

Uh we give a deadline to complete the training, we send out multiple reminders to staff who have not completed the training.

We'll send out reminders even after the deadline passes, continuing to send reminders along those lines.

We part of the updates to the chancellor's regulation A20 was to require supervisors to make sure that their uh supervised employees are completing the training.

And that when were those updates uh to the chancellor's regulations made?

Those were made in May of last year.

So the only change between 2024 when this report was released and they said that this was an issue, and now is that you told supervisors to tell their employees to complete the training?

Is that I'm sure that was a change in the regulation, but we're continuing to try to increase our training completion percentage through updating awareness guidance.

We send out principals digest messages, superintendent guidance, superintendent digest messages.

Um so we're we're constantly trying to to make sure that our staff are aware of their obligations and that they're meeting those obligations.

It sounds like a huge it just sounds like a huge gap, and this is a report from two years ago.

So yeah, sure.

Let me jump in.

Please.

And I want to say that, you know, as somebody who supervises our superintendents and therefore our school-based staff, we take full responsibility to ensure that those adults who are working with personal identifiable information, we are doing everything in our power to make sure they're getting that training.

And that is the priority staff, and we will work with our team internally to make sure that that gets done.

Okay, do you do you have goals for yourself?

I mean, we can call them smart goals, whatever acting you want to use, but you have you have goal, you're at like eighty-something percent now.

It was 73% two years ago.

Do you have uh goals that you're setting for yourself in terms of ensuring that this as we know AI is changing rapidly, it's getting into the classrooms.

More and more vendors.

I mean, this is important.

So what are the goals that we've that the DOA set for themselves to ensure that it is actually a requirement, and to me that means 100% compliance?

Yeah, I I think the goal is 100% completion.

I mean, that's that's what we're striving for.

It's it's and we recognize the difficulty in getting to that number, but we're we're constantly trying to get to that number.

Okay, I I I'm I want to move on to my other questions, but I am interested in hearing what steps since 2024, what changes you've made besides saying something different in the chancellor's regulations and sending out more emails and principals' digest is being done to do this uh critical training that was identified in the controller's report.

Uh the audit.

Uh the audit also identified weakness weaknesses in DUE's technical controls that were not disclosed publicly because they were confidential, correct?

Yes, that's right.

So is the DOE in a position today to share more information about those weaknesses and what it's done to remediate them?

I don't believe we're in the position to disclose that.

I think part of the reason why that uh portion of the audit report was kept confidential is because there's information in there that could be potentially could potentially jeopardize our uh information technology assets and our and our security posture.

Like your mom's made a name last four of your social and uh first pet's name?

Something like that.

Something like that.

But what that means to me also, um, I'll say the audit states the weaknesses involved system monitoring, unsupported systems and firewalls, correct?

Yes, that's right.

And have any changes been made to your systems to improve system monitoring uh and firewalls and other unsupported systems?

I could defer to to DIT on that.

I I as far as I know that work is is underway.

The answer is yes.

The changes have been made or the or or we're talking about doing the work?

Both.

Tell me more.

Uh I I don't really want to go too far past what has already been disclosed with the areas uh and this is this is common practice in a security, the security side of an audit that some of the information is not to close, just goes publicly, uh, but it was disclosed to us, and to the extent that we said yes, we agree and and could make the technical changes to improve security.

Uh we've either plan to do that or have done it.

So this again, this is a report I just from 2024, and I there and there in a confidential report you just shared that there were specific recommendations made in the confidential report in those two years, in addition to whatever changes that you are implementing outside of the recommendations in those two years.

Have you implemented the recommendations made by the controller's report?

So sorry, can I just enter it?

Please, what what is the 2024 data you're referring to?

I have a report for uh the data was from 2024 in the report.

I apologize.

So how many of those please thank you?

Let me uh I'll try to uh parse this because I would like to be able to answer the question.

Yes, thank you.

Let's say let's say there were five or six recommendations.

Okay, not disclosing specifically what they were.

I think it would be appropriate for us to look at those and provide back as much information as we can about the status of each one of those.

Yeah.

And I think I I mean this is all part of uh a pattern, right?

I mean, everyone from the DOE and the council, I think we care a lot about the same things, but what we're interested in is interested in is the implementation of a policy that keeps our kids safe and keeps our kids learning.

So I I um I want to turn over second round to to council member Wong.

Thank you, Chair.

Thank you, Chairs.

Um based upon what I hear today, um, AI 2s are already being used with inside DOE either by administrators or like principals and teachers.

Uh can you tell me uh which tools are uh being used right now and was there a vetting process on the the right uh like how did you choose these AI tools?

Please answer that.

Thank you.

Uh I want to answer it in a few ways.

Uh all of the tools, the digital tools that we are that we have we have a long list on our website of a lot of digital tools that have been through um the ERMA process, which uh Scott could probably speak to where to find it.

This in terms of what's actually being used in schools, we are working currently with our principals and superintendents to gather that information.

Our current guidance is where we struck the the fine line that this must be going, must have been approved through our ERMAR process and adhere to all of our privacy trainings and uh regulations.

Um and so at the adult level, there's um, you know, there's tools that adults in our in our charge have adopted, are using for teacher-facing, leader-facing uses.

Um, and then we are also gathering what are the top all of the different applications that are being used in our schools at the student facing.

We're doing that now.

Um but as we're collecting that, I just want to be clear there is a bottom line expectation that these whatever is in use, um, is through the Irma process.

Okay.

And uh, if not.

Oh, wait, let me be really specific.

Are you using like with by teachers, principles, administrators?

Are you using chat GBT?

I'm sure some are.

Okay.

What about perplexity?

Perplexity.

Not widely used.

Okay.

How about Claude?

C O A U D, no?

Not widely used.

How about Gemini?

Yes.

How about consensus?

Okay.

How about co-pilot?

Yes.

Okay.

Um what do you have to negotiate licenses for them or you just open it up and start using?

Because in my understanding that you may need commercial licenses for for the site site use of these AI engines.

Can you talk about that?

Sure.

Maybe our two largest some largest providers like Microsoft and Google have free versions of their copilot and Gemini, and also got paid versions that have expanded capabilities.

Uh in both cases, those products have been approved, and central offices and schools can purchase the advanced uh features and staff can use the free features.

And that is based on existing licenses you have with Microsoft and Google?

Correct.

But no no one like really looked into the engine.

Like, is this is this giving me the correct results?

So I mean, um, anyone ever looked at it because the quality of these AI engines vary?

Anyone looked at uh look at that, you know, and say Gemini may give us better results than chat GBT, you know, is like comparing.

I mean, our full expectation is that these are very limited use cases, um uh and that human judgment always has to be a core part of this, um, uh particularly when it comes to instruction.

We want to make sure that it is always aligned to the curriculum and the standards that we expect, and that there are much more trusted sources that educators and teachers should be using for those purposes.

We do know that there are some cases where administratively um for efficiency sakes um uh where they can be useful, but even in those cases, it is absolutely critical that all the adults that are using those tools are fact-checking and ensuring that everything that is being produced is accurate.

Okay, thank you.

Thank you, Chair.

Thank you.

Thank you, Councilmember.

Council Member Hanif.

Thank you, Chair.

I had asked about whether public schools were collecting uh biometric uh data, and while the schools may not be retaining that information, it's my understanding that Amira, which is being used for reading support, is collecting biometric data.

Um, is that something you all are you all know about?

What was the conversation with Amira around uh the storage of data specifically?

Oh, thank you for your question.

This is I'm gonna turn it in over to Tara, um, but I really appreciate you giving us also the opportunity to share how that is used.

Absolutely, thank you.

I mean, I think that you know I reading is is one place where this technology is, I think, going to be a lifeline.

Um, but in the same vein, if student voices are being collected, and we're not sure where that might be going to, who it is being used by aside from Amira, that is a problem.

Thank you for your question, Tara Carroza.

Um, so Amira is using student voice in an interactive way with their AI agent in order to provide um reading fluency practice feedback targeted towards that that individual students needs.

Um, in terms of our data processing agreement with Amira and with all of our vendors, there's a one-year data, and I'm I'm gonna pass it over to Dennis, but I'll share, you know, as I understand it.

Um after one year, all data has to be deleted, it has to be hosted within the US, and there is no use of uh student PII or biometric data that goes out to any other large large language model used to train any other models used to be monetized.

So using any form of student data is not allowed to go out based on our data processing agreement.

And that uh agreement can can that be shared?

Uh I'm gonna pass it to Dennis as the chief data privacy officer.

Yeah, we we generally don't share our data processing agreements uh publicly.

Why is that?

We it it's just uh like any other internal DUE document, we don't share.

I mean it's really hard to then trust what's happening.

There is a portion of the uh agreement that is posted publicly, it's the supplemental questionnaire.

Does it include this piece about the deletion after one year?

And there, yes, it does speak to how the vendor is.

I would love to just be pointed to where that is.

Um I also want to know on on the uh reading comprehension piece.

Uh has there been a sort of evaluation to understand if reading comprehension is improving?

I mean, I know that if I was using this in the fourth and fifth grades, I would be a much better a much stronger reader.

I I struggled with reading, and I think voices like hearing someone else speak to me and would be so helpful.

Um so I'm also curious how are you evaluating how this technology is uh delivering?

So currently, um, due to the do the due to the new nature of Gen AI empowered tools like Amira, we don't have that longitudinal data um as of right now.

We have it in lower grades, uh, that we can share the current impact of NY in NYC public schools, but part of this tool evaluation framework and part of our research approach moving forward and understanding how tools are being used within our local communities within our urban school district and the rich uh cultural uh diversity of New York City public schools, we need to do that work.

Again, we need to see what the vendor says.

This particular vendor does have a very large research base.

They've done their own RCTs, and what do they what's their kind of uh like what do they what do they say that a student can learn reading by or uh you mean like a grade level?

Yeah, like I can for anybody who's for the students who are uh using this tool to strengthen their reading comprehension to like by what stage of the of the school year um are they meeting that threshold?

I'll I can follow up with the exact data, but I'm gonna tell you what I remember off the top of my head.

Um, in terms of students, it was it was students using it uh up to three times a week for 10 minutes a day showed about a four to five month acceleration in um reading comprehension improvement compared to students who are not using it, um, and that's for last school year.

So we can share that exact data.

That would be airpoints.

Um, so you can see that.

Again, that's in the earlier, that's K-2, where you know we're looking at, we're looking at screen time, we're looking at what use, but Amir is not a student sitting in front of a computer and just interacting.

It's within the instructional program of a class, it's within the instructional program and high quality instructional materials at grade level of our New York City reads program, and it's a supplemental tool.

It's not that's not the core instruction.

The teacher is the core instruction.

Um, when I visited schools that are using it, it's in small group, it's extremely limited, and it's not uh just one-to-one for an entire class period.

So I do want to click on that.

I think that's fantastic.

And if there are other data sets that that you all have, which I hope you're keeping track of in terms of like curriculum uh and retent information retention or improvement of a specific skill, I would love to get a hold of that.

Okay.

We'll follow up with what we have.

Thank you.

Thank you, Councilmember.

Uh, one quick question for Mr.

Doyle.

I just wanted to clarify, you said you don't uh follow up on Council Member Hanif's questions.

You don't typically share the data processing agreement.

Yes, that's right.

Is that a you don't typically do it?

You can't or the DOE won't.

I think it's as I understand it, DUE policy, we don't make publicly available or publicly post um like internal DUE records or DUE contracts.

But it's not that's a policy that's a choice, and so it would be a choice to share that information with the council, correct?

I would have to defer probably to our records access unit on that question, and they handle the the you know disclosure of records.

Okay.

I imagine it's okay because you didn't sign an NDA according to testimony, so I'm sure it's fine.

So I think we can expect at least the council um to have those those data processing agreements because this also came up, um, you know, uh a lot of concerns about uh this uh particular program uh using Claude using using other platforms to process its data, which leads to whole other mess of issues.

Um so we look forward to seeing those agreements.

Uh the council looks forward to seeing those agreements.

Uh council member Zhuang.

Thank you, Cheers.

Uh, I have some question about my bill for disability.

Thank you for coming to testify.

Um I just want to know do you have the number of kids in different zip codes and a different borough needs, special needs kits?

How many numbers, Maggie, by zip code, by borough, by race, and by language access?

So we do have uh, I believe most, if not all of that data that's publicly reported in city council special education reports.

Um, and uh we can I don't have them at the top of my head, uh, but we can certainly uh go over with you uh that data at a you know at a time outside of this hearing if that's if that's helpful.

Okay, and also by race, um I look at the data is 75% students disability in public school system are black and the brown uh black or Latino, and also um for the private school tuition reimbursed went to black and the Latino students, only 24%.

Is that correct?

And uh most of tuition reimbursed 71% are going to white students.

Is this data correct?

So, in terms of the students who uh by uh demographic ethnicity in public school settings, um I think you said 75% are black?

Black and the or Latino.

A little bit closer to 50 percent.

Uh a little bit closer to that.

However, I do want to acknowledge your your question and your comment and and be very clear that the division of inclusive accessible learning, which is a division I work in, but New York City Public Schools at large and certainly this may oral administration are very focused on equity and integration of public schools.

Our chancellor has said very clearly uh about true integration and that include students with disabilities, so this is uh something that is is very, very paramount and something that we are working on across the board as well.

So what's for the last year, City has been reimbursed for the spatial needs kids' education more than 72 million dollars.

71% went to white family, and but in York City system is 12.5 of students with a disability in York City schools are white.

What is city doing ensure students with a disability who are not white get the help they need?

Thank you for asking that.

Um that is a focus of our work, and uh some just a few of the things that we have done as a city.

Uh one I want to share that we have opened uh fully inclusive districts for autistic children in neighborhoods and communities that were historically under-resourced and unserved.

Uh and in those districts, every student who's coming into kindergarten receives an uh seat at a high quality program close to their home, uh, one of our specialized programs called AIMS uh or Horizon.

And those were intentionally opened uh in districts 5, 12, and fourteen just to start because those districts often don't get their sometimes districts that don't get resources needed.

We have also hired and trained uh over 900 in intervention teachers who focus on literacy and reading for children, and those teachers work primarily with students who have historically been marginalized black and brown students that is a primary focus in their work as well uh another thing that we have done is built out our assistive technology there used to be certain neighborhoods where you saw a lot of assistive technology requests and then other neighborhoods where you saw very few and we've intentionally done training and outreach to families and teachers and school staff in those districts and we've seen those numbers really rise so leveling the playing field around uh literacy assistive technology for our students and building programs in communities that historically have been under resourced all of that said council member we certainly have work to do uh and I'm excited about our work with first deputy chancellor junta and her team on this kind of equity and inclusion for all students moving forward we know that we have a what's called an overreferral problem of black and brown specifically boys being overreferred for special education we're really clear on that and that is something that is paramount training our school psychologists training our school leaders and working through that so that the right children get the right services the right way where and how they need them not based on race ethnicity or uh family income so lastly I just want to share one other thing uh we opened 80 called sensory gyms seed sites across New York City in the past before we we opened these four years ago we've continued to make more and more we started with 10 this year we ended with 79 pro before that only families who had the money to pay privately or had the opportunity to sue New York City public schools got that kind of service and support what we did is we built 79 of these that are free of charge in neighborhoods so families can access that for free and not have to sue or pay privately so we are working at this very important issue from a lot of different ways.

I have one more question.

And also I represent a large immigrant community a lot of the time parents come to my office saying hey I have to travel one and a half hour to bring my kids to Queens to go to a spatial needs school and do you have data to show um how many kind of kids in York City have this type of situation?

Yes we we do have data on students who are enrolled in district 75 which is the citywide district for children with uh more some intensive disabilities and very often those children are going to school far from home because the program is not located in their neighborhood and what what we are building and the three districts that I mentioned and we are expanding our specialized programs to be close to home so children don't have to take the bus get up at 7 a.m and take the bus for an hour but we do have data on those programs as well as students currently enrolled in District 75.

Do you also have the data for the family because I represent large immigrant family a lot of family doesn't speak English at all from the data I see research data not DOE data saying wow five kids family speaking English as a second language of spatial needs kids do not have service is that correct that is not the data that I have and around one in five families not received who are not English speakers not receiving services so data said never receive services never we that if those are children who've been evaluated for service and never receiving it, I would very much want to look into that because uh that's not something that that where that we see in the data set that we have.

And then because a lot of the time the family doesn't speak English, they don't know how to access the data, even they do not know where to report this situation.

Yeah, and and we have there's there's a lot of resources that we have for families who do not speak English and or even families who are new to New York, how they can get services quickly for their child without even any documentation.

Um, and we would love to have conversation with you, any other council members on what we should do more, better to get that word out to families because those services do exist, and we want families to feel that they can ask.

And sometimes we hear that families are fearful to ask based on sometimes immigration status, or they're if they are not comfortable with the English language, and we want to do everything possible.

Uh, and if we have to do, we we certainly would be willing and happy to do more to reach more families.

One thing I do I do want to say, since we have uh the the first deputy chancellor and her team here is that the first stop for any family is their school.

And if a family is not receiving service support that they need, the first place that they should go and be supported is in their child's school as well.

And the young 2024 September on the hearing, we already said auto borrow bus routes for students with disability usually two-hour away, two hour one-way, four-hour every day.

When this type of situation happens, um city permits bus routes up to two hours each day.

Importal routes without IEP mandated travel time is one, a half hour.

Uh, is any way we can fix this issue?

So, in terms of the bossing issue.

Um, and I know you're you mentioned children with and without IEP.

I will just say that as a system, the best way for us to fix that issue is to continue to open more specialized programs for students with disabilities in their home and local communities.

That's something we've been advocating for.

The uh uh chair Dinowitz has heard this uh very often from us, and uh we know these programs work, and that is our best way to resolve part of the busing issue is not to put children with disabilities on buses to travel an hour from their home.

Yeah, if we have all the service in that neighborhood, then it will be much easier.

If we can have the data, it's easier too.

Thank you.

We will share that district seventy five uh data with you.

Thank you.

Thank you, Councilmember.

And I want to personally thank you because I think you know how important it is to support our students with disabilities, and I just appreciate your legislation on this, and uh I will remind everyone according to testimony during our executive budget hearings that I think we're all in agreement that we do need to open up more specialized programs for students and IEPs, and there's a little roadblock um called OMB.

Uh that was that was the testimony from the last hearing.

Uh I I want to go back to the audit, but before that, I just want to go to the tool review, the AI tool review.

I'm a little confused by the guidance that currently exists on your website.

So if you could walk us through that would be very helpful.

The current guidance says the enterprise request management application, Irma, is DOE's privacy and security review process.

The guidance also says that AI tools can't be used with any student or staff data unless they've passed the IRMA process, correct?

Okay, that for the record they nodded their heads yes.

Oh, sorry, yes.

No, it's yes.

That just means you're good listeners.

That's all it means.

But the guidance says Irma does not evaluate algorithmic bias, equity impact, or instructional effectiveness, correct?

Correct.

But you also say on another part of the website that the legal team reviews bias.

Does the legal team review bias?

No.

I don't that may be a misprint on the website.

I'm not sure what part of the website is.

You should have had AI fact check your website.

Possibly.

I'm just kidding.

Relax, everyone.

I just want to read it to you so you so we're all on the same page.

Legal and compliance review.

New York City public school teams review legal terms, privacy protection, security measures, instructional value, and AI-specific issues like bias and transparency.

So you're testifying that that's a typo.

Well, say transparency is something that gets conducted as part of a legal review, but that's not the same as bias.

I mean, transparency.

The word bias is there.

So you're testifying that that is a type of.

The legal team does not review for bias.

Okay, so it's on the website, so I'm sure your webmaster will go and delete that right now.

But this is this is the guidance, by the way, that's been up for however long for weeks or or since since it was it was it March it was published?

It's been up for months.

Uh so please correct that.

And I also want to add so Irma does not evaluate algorithmic bias.

We've determined the legal and compliant legal team does not review bias either, and you'll fix the website.

And New York City public school is actively building the capacity to review for algorithmic bias.

That is correct that that is something the public school team is working on.

Say the question one more time.

Is the website accurate that New York City public schools are actively building the capacity to review for algorithmic bias, which yeah.

Go ahead, Scott.

I guess I'll start.

Um for tools that otherwise look like that they have a great educational purpose or maybe even operational support purpose.

I think we want to extend our review of that to bias.

We can't do all tools all the time, but for the ones that we want to use and encourage our people to use, we want to make sure that the bias, the bias is also evaluated.

There's a variety of ways to do that.

One of the best ways is to make sure the tool uses a curated data set to do its work.

Uh and we have several tools that follow that kind of uh processing pattern.

But you you don't have a policy that all tools that you don't have a requirement that all tools use a curated data set or or just the data that you that you allow it to use, correct?

That's what I require.

Yes.

With a large language model, you know, takes from here.

I'm saying that is the New York City public school currently building the capacity to review for algorithmic bias and all of their approved tools and all the tools they plan to approve.

I think I just located the section that you're talking about in the guidance.

And I just wanted to ask uh Dr.

Pete if she can come up.

I I actually just quickly sent.

I think this the thing that you see, and um I just wanted to know if Dr.

Pete can speak to it.

Yeah, I'll read it to you.

Bias and equity review.

Irma current currently evaluates data privacy.

New York City public school system is actively building the capacity to also review for algorithmic bias, equity impact, and instructional effectiveness.

But I'm focused on that bias and equity impact piece.

Sure.

So we started with the instructional effectiveness with the tool evaluation framework.

Um, the equity impact that we're looking at, particularly particularly in tools, we started with exploring that through our equity and uh ethics and equity fellowship.

I want to make sure I get this that naming right.

Um, a teacher, uh, a teacher training uh fellowship over the past year in terms of bias and equity impact in tools, um, given the amount of tools and even across the private sector, there are not bias uh evaluation of tools readily available, and so as as the New York City public school system, we're looking at how we can actually work with our existing big partners on the enterprise level, and I'm looking at our partners at DIT as well with the directionality of OTI and what the city's gonna do with bias evaluation and equity impact because it's a new area, just like the research um within within AI.

But but algorithmic bias is nowhere near a new area.

I mean it's it's it's in AI now.

You're asking.

But is is it sounds like you're going to be having conversations?

Your website says you're you're building the capacity, but it sounds like you're kind of talking about going to be working with enterprise partners to figure out what maybe how to figure out what bias looks like.

That's kind of what I'm hearing.

I understand that's what you're hearing, and I I respect that as well.

And I hear that.

So let me try and clarify.

Thank you very much.

Sure.

Um in terms of tool evaluation at our size and scale, by algorithmic bias is something that we can evaluate for.

Doing it at the scale and across the number of tools we have in New York City public schools is something that we have to do in development.

It's not something that we have for you right at this moment.

It is something that we are thinking through, we're working through, and looking at what's the availability of us in in terms of what tools are available to actually evaluate that the algorithmic bias within all of our tools at the various levels.

We have enterprise tools, we have smaller ed tech tools, we have curriculum supplemental tools.

So it's across a wide number of tools.

Who's so whose decision in the DOE is it to approve tools in the first place without discussing or figuring out what the bias or the equity impact is?

The first step of any tool approval is for a principal, a superintendent, or a central leader to demonstrate to demonstrate its educational value and impact.

That's step number one within the IRMA process.

So it has to be an educationally purpose-built tool.

That's number one.

I'm gonna I'm gonna pass it on the bias part, not because I'm passing the question.

I'm passing it to Dennis in terms of what our what our current expectations and our current um the current IRMA process has in terms of what vendors can share to us.

It's not a requirement, it's not a requirement within our process at the moment, but it is something that we're looking to build into it.

So I'm gonna pass that to Dennis.

Yeah, thanks, Tara.

Um right now again, like the legal, the legal team at least is not evaluate for bias.

I I'm not exactly sure which part of the website you were referring to, but if if you get it to us, we could try to correct that because I know there's other parts of the web of the guidance and website that um are clear that that the IRMA process is not yet evaluated for bias.

Um and I think when we're talking about tool approval also, we're talking about approval just from a data privacy and security compliance standpoint, which is really just the the minimum floor for a tool to be used with student data.

I think in order for an AI tool to be approved, like it has to go through additional layers of approval beyond that.

Irma is just evaluating for the data privacy and security possible.

I understand.

But given the explosion of AI and the challenges we face, can the DOE commit to ensuring that bias and an equity impact are part of the tools that you use to evaluate these tools, which uh you've already released into the classroom, and or at least to our schools, but can can that be one of the things you commit to doing?

So let me get to the answer first.

So I just be I have a printout of the guidance.

So yes.

So the section that you're referring to is under gaining clarity together, concerns and considerations.

Earlier in the document, there's a section three, how tools are evaluated, and we clearly say there that it uh the ERMA process currently reviews tools for data privacy and security.

It does not yet evaluate for algorithmic, today's not my day with this word.

Uh bias actually instructional.

It is, it really is.

Um, terrible words.

I just wanted to show the nuance of this is how our tools are evaluated in section three, and that that is under the considerations and sort of next steps uh for that.

And in terms of like bias and tools, you know, but you know, bias lives um well beyond our tools as well.

It's uh part of our commitment around um ensuring a culturally responsive and sustaining curriculum.

It's been part of our investment, even in the hidden voices, the work behind the integration in our curriculum, because we also know that just our curriculum, you know, we're always checking is it representative, is it biased?

What other you know primary source documents can we bring into a curriculum?

So in terms of the ask on the AI tools, I'm just gonna take that as um under consideration.

I don't I can I commit to even the capability to do that as part of what we hope to, you know, consider.

I just don't want to lock in on the feasibility of how we would be doing that yet.

Okay, I mean I think you should.

I mean, I think there are certain baselines and and I think what was testified to earlier, it seems as if there's a very minimum standard for approving tools that um you know that can pose a danger cognitively, emotionally, academically to our students, and and that there aren't certain that we're not taking this a little a little more seriously in terms of like the you know the minimum standards is is a little shocking, honestly.

You know, the that we are putting the car before the horse by releasing these tools with the I don't know, the same guidance it seems that we're using for any other any other vendor, but this is a new world.

And we have to really take seriously all of the elements that have been brought up today, including bias um and equity.

And I and I do want to quote because you because um Mr.

Doyle asked where it was in the ARMA process, step six in the ARM, according to the website, and I understand it's a typo, but I did want to cite it for you.

Step six, legal and compliance review.

The teams review legal terms, privacy protection, security measures, instructional issue, instructional values, and AI specific issues like bias and transparency.

Number six.

I want to end with um just more about the state controllers report.

Um about data breaches.

Okay, I I think very concerning to all of us, correct?

Yes, absolutely.

Good answer.

Oh, we're nodding again.

Okay, yeah.

That's called saying yes out loud is called implementing feedback.

We love to love to see that.

Um and of course the DOE thinks it's important to report data breaches in a timely manner when they occur.

Um, apart from being important.

Reporting the breaches is required by part 121 of the State Education Department's regulations.

It requires the DOE to notify the State Education Department's chief privacy officer within 10 calendar days of discovering a breach.

The controller's audit reviewed 141 breaches and found that the DOE reported 67 of those breaches, which is almost half of them late.

Explain why the DOE failed to comply with the law in almost half of the instances that the controller examined.

Yeah, thank you for the question.

Um I think as far as reporting to the state goes uh when we have a breach, I I alluded to this before in another answer, but part of our difficulty is that we rely on our either our vendors to come to us and report this information to us in a timely fashion, which unfortunately does not always happen, or we we rely on our staff to come to us and report the breach on or unauthorized disclosure to us as soon as possible.

And I think one of the ways that we've improved on that front, and I think the the figures in the controller's audits show that at least we're trending in the right direction, is by uh improving our awareness campaigns, improving our training.

I understand there are we are not 100% completion for the training, but even that number is improving.

We're finding that we're getting more um more staff that are aware of their privacy obligations, they're in contact with our office more and reporting more to us.

Um so I think we're improving on that front, and it's a it's a tight timeline, I think 10 days to report to the state.

Um so those are some of the factors that go into it, but um, yeah, we're we're moving in the right direction on that front.

And now you so you've spoken about how you ensure or are attempting to ensure um that professionals uh get the training so that they report, but what about the vendors?

Our vendors are required under our data processing agreements to notify us as soon as they have um confirmed an unauthorized disclosure or a data breach.

Um there are sometimes difficulties with getting that information from vendors in a timely fashion, notwithstanding uh what they're required to do under the DPA.

Sometimes vendors don't contact us at the right point of contact.

They will, for example, contact a school and not contact the student privacy office so we can affect the notification right away.

But that's not that I'm sorry, but that's not in their contract that in the instance of a data breach contact.

It is.

It is, but we said we still have difficulty with vendors, you know, complying with that part of the contract, and we're looking into uh, you know, uh different ways that we can strengthen our DPAs to make sure that vendors are going to be held accountable to this to make sure that they're reporting it in a time.

So you drop them as a vendor?

Say that again.

You drop them as a vendor if they don't comply.

I didn't no, I didn't say that.

I know I'm asking.

Would we I mean we have we have you?

Have we before?

Would yeah, have you?

Would you?

Did you?

Yeah.

Yeah.

In the case of um Illuminate Education, for example, that was uh a data breach that happened about four years ago.

Um there was uh you know, a pretty wild wide scale uh breach that happened, and New York City public schools decided that we're no longer uh going to be doing business with that vendor.

Okay, good.

So that so they failed to report uh for two years, and then once you found out about the breach, how long did it take the DOE to report to the state the data breach?

I'm sorry, but what was the You just said that the two years?

How long did it take for Illuminate to report to you after the data breach that there was a data breach?

It took I think what happened with Illuminate and this precedes my tenure in my current role.

Um but Illuminate informed us of an incident, but they didn't completely categorize it as a data breach until several weeks afterwards.

Um I think it was a few months before they officially deemed it a data breach.

Um, and so it was a matter of a few months from the time of the incident to the time that we were officially notified of a data breach.

Well, in the Illuminate incident, 380,000 families uh were not informed.

This is what the the uh contract you brought up, 380,000 families were not informed until May of 2024, more than two years after the incident, and seven months after Illuminate had informed the DOE of the additional students involved according to the DOE website.

Delays in informing families can have devastating consequences because children are top target for identity theft since they don't have credit ratings.

What's the explanation?

I mean, these are inordinate delays.

So what I'm asking about is the the delays with the DOE.

What what explains the delays between the DOE and the reporting of this information according to the controller's report?

Well, for the Illuminate case, uh there was actually so there's the initial incident.

We eventually received the information necessary to make the notifications, which is often part of the difficulty and the delay in getting the notifications out as we rely on the vendors to get us that data as soon as possible.

That can also that can often take months.

Forensic investigations have to be completed before we get that information.

As soon as we got that information, which was about four or five months after the initial incident, we sent out the notifications.

I think it was about a year and a half later, uh, Illuminate came back to us and informed us that they had actually discovered that additional students had been impacted, and that had they they had not initially informed us of.

Again, we had to go through that process of obtaining the data from them, uh, the whole forensic investigation, and you know, go through our process of matching that with our own records so we can send out notifications, translate letters, all that, get that done as soon as we could.

Yeah, the controller found the deal we failed to report breaches to families within the required 60-day time frame.

Um obviously parents cannot act to protect their children's information or work to mitigate damage from a breach if they're not timely, uh if they're not told in a timely fashion that their child's information was was disclosed.

So what steps of the DOE has the DOE taken to fix these lapses?

Yeah, that's a good question too.

I mean, I think again, a lot of this goes back to our training and awareness programs, um, making sure that our staff are completing the training is is very important, and we're we're pushing to get that like I said before to a much higher number.

No, no, I know.

Sorry, I discussed when the DOE, not when like an individual teacher or school.

I'm talking about when DOE Central gets this sort of information.

Like that, you know, I think that's a good idea.

Unless I think that individual teacher or principal submits this information to the state.

I think we would consider the date of discovery to be the date that like the principal at the school learns of an incident, or and then there's often some sometimes there can be a gap between them reporting that back up to the student privacy office so that we can then turn around and send the notifications.

And when we do send the notifications, when it is a cybersecurity incident, like in the case of Illuminate, we also do extend free credit monitoring and identity theft protections to students, which is sorry, you do that or the vendor does?

We do that.

Do you well the vendor reimburses us for it?

Okay.

That's something that we have minimum.

I feel like there should be a penalty, but but it's we do get reimbursed for it, but it that's sort of another area where we try to go beyond, you know, what's required by Ed Law 2D to make sure that like our you know, once we do send those notifications that they have the these resources.

So, in addition to addition to the increased training or trying to get that number up to 100%, which is the requirement slash the goal, what is the DOE doing to fix the lapses?

So you've mentioned the training.

What other steps have they taken?

I think we're relying on timely reporting from all of our offices and all of our schools so that we can get this information out as soon as possible.

And there's no so you're saying there's no issue between once you get the data or information, there is no hurdle, no speed up, no hiccup for the DOE to then report it to the state.

Well, reporting to the state is one thing.

I think we're if we're talking about reporting.

And to families, yes.

If we're if we're talking about notifying families, that's a little bit more complicated because we have to get the impacted data, we have to understand who is impacted, we have to then get if it's a vendor that's involved, we have to get the data from the vendor, then match it against our own record so we have contact information, translate the notices.

You have a 60-day time frame.

Yes.

Right.

I'm simply asking if there are any additional steps based on the finding of the audit that you were doing to improve this to ensure that you are notifying families in the 60-day time frame.

Yeah, and I think one of one of the things is that we're looking into is and we're always looking into is ways to strengthen our data processing agreements when if we're talking about a data breach that's associated with a vendor to make sure that they are held accountable, that there could be penalties, for example, for not notifying us in a timely fashion.

Well, is that is that gonna be in your contracts then?

Because I think the penalties are completely appropriate for a company.

Besides reimbursing you for paying for families uh data monitoring, uh is that gonna be included in your contract, so there will be a penalty when there are data breaches?

It's something that we we are looking into for including into our data processing agreements, yeah.

Okay, I mean I think that's incredibly appropriate given the scale that we're talking about and the amount of especially with AI, the amount of student data that we're talking about.

So we're not just talking about personal identifying inf information that you guys carry.

We're talking about all the stuff that they're inputting into whatever chat bots or or or uh systems that they're using, which goes a lot deeper, knowing that especially since kids use I hope you're teaching them they shouldn't, but using that these chatbots as mental health counselors and using them as as friends and as and as buddies.

Yeah, and one other thing that I I should point out too is when we're talking about data breaches and and when the controllers report refers to data breaches, we're not only talking about data breaches that are derived from third party vendors under edlaw 2D, a data breach is any unauthorized disclosure.

So if we have um a DUE staff member that's sending an email to the wrong recipient and includes one student's personal identifiable information in that technically that's going to be considered a data breach, it's gonna trigger all the same obligations under EDLO 2D.

And I think a lot of the delays that we see are sometimes again the staff not reporting that to us in a timely fashion so that we can get the notifications out.

Okay, thank you.

I'm gonna turn it back over to Chair Della Rosab.

Thank you.

So when talking about uh it says here, paragraph four states that vendors complete a security questionnaire covering encryption, access control, and compliance standards.

Um New York City Public Schools security team reviews and approves it.

Can you clarify what compliance standards are referenced in this paragraph?

And uh who where are these standards issued?

Who are these standards issued by?

Are we you're referring to the data security review process?

Yes, I'm um referring to uh to the security check part of it.

Okay, I'll defer to to DIT for that question.

Sure.

So the review process that goes on within the ERMA process is three steps.

Uh the data processing review, which is legal and working with the vendor, getting the proper documentation, and then DIT uh does a review of the technology related to the the um system, and we get into those areas that you just described.

Uh we have security professionals that know how to do that.

Uh we have some tools that we also use to um support that process.

And the third step is a review from OTI Cloud, which uh also looks at the technology and its security and where things are hosted uh and make sure that it's secure from their perspective.

So, but who sets the standards?

Is it, you know, OTI?

Is it DOE?

Like who it's DOE.

Okay, okay.

And then I had another question.

Let me find it in here.

Um, here it is.

Uh I think Councilmember Hanif asked about biometrics, but we wanted to know a little bit more around behavioral monitoring, which uh it states that it's prohibited at this moment.

Behavioral monitoring is prohibited in the guidance.

Correct.

Red that's uh in the red uh area of terms of uh a feature of a system, correct?

Uh but the guidance also says that New York City Public Schools is reviewing how existing policies apply um to AI related collection and use of biometric and behavioral data.

So is it prohibited or are you all uh exploring or studying or uh considering biometric use for behavioral monitor?

So the biometric we talked about is it is illegal to to use uh behavioral monitoring again is red area.

So I think in both cases we're you might be continuing to look at that almost to be sure that if any tool comes that's begins to encroach on those areas, we point out that we can't allow that to happen.

Okay, so you're not considering it or yeah, we're not opening it up, we're making sure it stays closed.

Got that.

Okay, and then uh let's see.

Does the does the prohibition also apply to device monitoring, emotional recognition, facial recognition, profiling, or any other forms of surveillance?

Broadly, yes.

Okay, all right, thank you.

Yep.

We have um a number of follow-up questions that we'll submit to you in writing.

Um, I want to thank the administration for coming today.

This is obviously a very huge issue that I think everyone here uh is really concerned about.

Um we uncovered a few things.

I think penalties for these these these companies that are involved in data breaches is perfectly appropriate, and we look forward uh to seeing those contracts as was discussed in in the hearing.

Um there's a lot of work to do around tightening up reporting, um around and transparency around these tools, and that those are my two pieces of legislation that we look forward to discussing with you further.

And of course, as I started out with uh what it means for our students and the cognitive offloading, and Dr.

Chang mentioned um civics and digital literacy.

And we at this point in the city's history have an incredible opportunity with portrait of the graduate to make sure that is embedded.

You can decide, and we could decide together to make that a requirement to evaluate it to make sure it's happening and not wait for guidance from the state in order to require us to do it.

This is our city, this is our school system, and we have an obligation to give our kids the best future possible, and that includes all of the things that were mentioned today.

It includes protecting our children so that they have, you know, every opportunity that they deserve.

So I want to thank you once again for coming.

Uh, we'll get started on uh public testimony in just a moment.

Thank you.

Thank you very much.

No one else.

Okay.

Thank you.

We'll begin public testimony in a moment if you'd all take your seats and your conversations outside.

I now open the hearing for public testimony.

My co-chair and I remind members of the public that this is a government proceeding and that decorum shall be observed at all times.

As such, members of the public shall remain silent at all times.

The witness table is reserved for people who wish to testify.

No video recording or photography is allowed from the witness table.

Further, members of the public may not present audio or video recordings as testimony, but may submit transcripts of such recordings to the sergeant at arms for inclusion in the hearing record.

If you wish to speak at today's hearing, please fill out an appearance slip with a sergeant at arms and wait to be recognized.

When recognized, you'll have two minutes to speak on today's oversight topic.

If you have a written statement or additional written testimony you wish to submit for the record, please provide a copy of that testimony to the Sergeant at Arms.

You may also email written testimony to testimony at council.gov within seventy-two hours of this hearing.

Audio and video recordings will not be accepted.

And I may remind you throughout the hearing, uh, adults, your your testimony again is limited to two minutes.

And if you go over, I will just ask you to please submit uh the rest of your testimony in writing.

Uh our first panel, I was like to call students.

Um our first panel is a student panel, Selena Wu, Ariana Ahmed Misha, Nelani Joseph, and Odin Adler.

You may begin when ready, and please state your name for the record.

Hello, good afternoon.

My name is Yelani Joseph.

I'm 16 years old, a sophomore from Brooklyn, New York, and an organizer with the NYCLU's Teen Activist Project.

Thank you for the opportunity to speak today.

When people talk about artificial intelligence in schools, I often hear one assumption that students either fully support it or don't care about its impact.

But many of us do care because we're the ones experiencing these changes in real time.

My concern is how quickly AI is being normalized in schools without enough accountability, transparency, or guidance on how it should support student learning.

Recently, my friend and I were reviewing a reproductive health slideshow for class and came across a sentence full of scrambled words, errors, and unclear information.

Our impression was that the teacher created the classroom contact with Notebook LM, an AI tool.

And we were shocked.

Even though we were eventually able to make sense of it, we shouldn't have had to spend time guessing what the curriculum was trying to say.

Students deserve educational materials that are clear, accurate, and understandable, especially when learning about subjects as important and personal as our own health.

And that experience reflects a larger issue.

So much that's something I've also heard echoed when talking to my peers about AI and education.

I've had one of my peers recently tell me, quote, although AI is helpful in some ways, students often use it to cheat instead of studying.

And when exams come around, students can feel lost because they don't have that sense of security that they did before.

And another student, my friend, shared with me that, quote, her school is shutting down, and they're opening up a school for AI focused in their space.

I love my school because everyone was able to get the attention they needed, and taking that away feels like a step in the wrong direction.

Even though that plan for an AI school was paused for now, what stood out to me wasn't the opposition to technology.

It was a pattern.

AI is being normalized faster than students are being taught how to engage with it responsibly.

AI literacy has to mean something other than filling classrooms with AI products.

It has to mean understanding how and when to question these products and protect against the threat that they pose to teaching and learning.

Students are asking for accountability, clear standards that protect student learning, and standards that ensure human judgment stays at the center of education.

Thank you.

Thank you, Yelani.

Time review reset.

You start the you should worry about you, go ahead and testify, they'll figure it out.

Hello, City Council, and thank you for spending time out of your busy day to listen to what I'm here to say.

My name is Odin Adeller, and I'm a leader at the New York Civil Liberties Union's Teen Activist Project.

I'm eighteen years old and have just graduated Urban Academy Laboratory High School this week.

I'm here to talk about my views on AI and tech products in schools.

At my high school, we did a lot of writing, and before I went to Urban Academy, I didn't feel confident about my writing skills.

We're also not allowed to use AI to write our papers at all.

That was not a problem.

Without AI, I could do something else.

Ask my teachers for help.

And now feel much more prepared for college.

In college, which I'll be starting this summer, I'm sure there'll be much to write and it'll be writing that I'm sure I'll be expected to come up with ideas for myself.

Fortunately, I believe I can now put some of the many ideas I think about on the paper quite well without AI.

That also helped prepare for my high school's frequent class discussions where subject knowledge and critical thinking is essential.

Knowing how to write well-without AI also helps creatively.

I'm not sure I would have had as much fun or learned as much my playwriting class where I wrote up a play that got acted out to my school by just use AI.

My classes felt much richer due to me and that is avoiding using AI to replace thinking.

Unfortunately, I've heard that official guidance on artificial intelligence is being sent to schools, which uh encourages students um to use AI for research, exploration, and creative projects.

This is unwise, I believe.

I'm concerned that we'll guide those in schools how to use software that is not useful in the classrooms and schools, I believe not just the final typing up of a paper, but more, including gathering ideas, should all be done without AI.

At school, students should be taught to think for themselves.

I'm sure it's possible people knew how to do so for a while before AI.

That is why I'm glad my high school is as restrictive of AI as it is, for both the writing a student's produce and the creation of the material used to probably teach students to be critical thinkers, support that kind of AI restrictive policy, Jeff the City.

I'm also concerned about the risk of students.

They added privacy, it's required by the government for minors to attend schools.

Students should not be effectively forced to sign with digital rights to big tech companies to fulfill that requirement.

Instead of spending much taxpayer money on AI and data captioning products, which are not what students need, the city should focus resources on actual human education.

Thank you.

Thank you, Odin.

Thank you so much.

Hello, Chair Dinowitz and uh Chair De La Rosa and members and staff of the committees on education and technology.

My name is Arena Amid Misha.

I'm a rising junior at the Institute for Collaborative Education and a member of the Teen Activist Project at Night Blue.

This spring I helped plan a weekend event with hundreds of students to listen to their concerns about AI guidance.

And there was a very specific session about AI in that entire ceremony, and that conversation about AI went so far and so much that it went over time.

And even after the session, we could not stop talking about how much we absolutely hate that policy idea.

And critical thinking is the whole purpose of our education.

That's why we get degrees and everything, but AI is completely taking our skill of critical thinking because they're not letting us work.

So what's the purpose of education?

If AI is helping us to critical think in some ways, and also as students, we want real human teachers to teach us.

And we want real human teachers to give us feedbacks about how to improve our work, not AI.

And also by reading our hardworked assignments, because we worked hard for all those assignments every single day.

And to get to know how we think and how we do our works as a human student.

So as a human student, we want human teachers.

However, for many students today, they instead use AI bots, which tells them how to improve their works.

And this AI bots are so insanely perfectionist, they keep saying the students again and again to fix their work again and again until infinity, and this makes the students lose their it lose their all the confidence over their own work, which is really very bad.

AI policy at the end of the day pushes us into a fake and unrealistic world.

We do not want a robotic policy in this already robotic world.

And also for this reason, I want to have a hundreds of students who care would like to request you to push DOE to make a policy dependent on humans rather than AI.

Because rearranging a system after AOI, AI, is AI is gonna destroy it, is gonna be possible.

But the lives which is gonna be destroyed after this process cannot ever be fixed.

Thank you so much.

You time that really well.

That's perfect.

Thank you.

Hello, City Council.

Thank you for allowing me to speak today.

My name is Lena Wu, and I'm a rising senior from Francis Lewis High School and a member of the Teen Activist Project.

The feline market of AI is growing at a rapid rate, which is why it starts getting implemented everywhere.

First handedly, I know how strong artificial intelligence can be.

In fact, I've spent my past summers researching and building machine learning models at schools like Princeton and MIT.

This is why I would like to oppose the use of AI technology in a school system.

My comments will focus on the use of Gen AI products.

Firstly, the highlights of learning are the interactions and jokes my teachers make when teaching.

In my AP LAN class, we start off every day spending 10 minutes discussing different things, whether it is political, philosophical, or controversial.

These interactions not only creates a bond between the class, but makes the classroom more lively, and it sharpens our critical thinking.

This actually makes me look forward to learning and attending class, which is something that not a lot of teachers and students can experience.

Having to think about complicated topics and shaping them into concrete words are a skill needed for the AP LAN curriculum as well as the test.

Offloading tasks to AI jeopardizes these human relationships and means me, my classmates, and my teachers understand one each other less.

Furthermore, when Gen AI is used in schools, it is often incorrect.

Last year, I had an English teacher that graded our essays with AI.

Not only did his AI incorrectly deduct points from us, it also accused multiple of my classmates from using AI to write their essays.

However, we wrote these essays in class on paper with a prompt that was given to us on the day of.

There was no possible way for any of the us to use AI.

And we were unfairly given points getting deducted, despite following the rubric.

Compared to my AP LAN class this year, there's a huge contrast between my learning and relationship with a teacher that uses AI compared to one that doesn't.

Secondly, I have seen that implementing AI and robots into the education system has made learning even more boring for some students.

The rise of AI chatbots like ChatGPT and Claude has encouraged students to cheat, plagiarize, and break numerous honor codes.

But this issue is much more complicated than adults can acknowledge.

Rather than implementing AI into schools and allowing students to further rely on such technology, I think that there should be a focus on creating an engaging learning environment that addresses why students use such technology in the first place.

By encouraging curiosity and making learning more interesting, students may feel more confident, less afraid, and find less of a need for AI.

Thank you.

Thank you so much.

I have I have a couple questions.

Any of you can answer this, but you know, I often I I think there's sort of like an assumption that you know kids will do whatever to like make the work as easy as possible, right?

If there's like a chatbot available, they're just asking, right?

Um is it is it your uh feeling that your views that you don't want AI in the classroom because you know how important critical thinking is.

That's been your general you want more transparency and accountability.

Do you think that's the view of most of your classmates, even the ones who use AI in the classroom?

Um I think like oh, to understand that issue.

A lot of time, as we are not adults, a lot of time it's not necessarily I'm not being weird, but not necessarily always important to do whatever we want to do.

Definitely it's gonna make our works easier, definitely it's gonna maybe make us less think about it.

But what's the whole purpose of education is the main question.

So if we do really want to learn or make ourselves smart or want to critical think, then we would definitely need a process of education where we need to think about something before answering, not necessarily a bot telling us or um doing works for us, because at the end of the day, that is definitely not good for us under any circumstances because first of all, it basically lacks all the opportunities for us to learn anything.

The whole purpose is learning.

So definitely there must be some students, of course, who wants, not a lot of people, because people who genuinely cares about their future or wants to learn would not follow such kind of things or would do.

But most of the students, whenever I talk with my peers, and as I also mentioned that we had a conference about talking about issues like that, people absolutely hate AI, no matter how much they uh, some people might say that they liked it, but they do not like it.

They understand that how this is also impacting the environment in many different ways.

Um, yeah, I believe also at my school is very uncommon.

I don't think I notice anyone using AI, uh AI, um especially um, which is in general, especially for like the writing and the art um that's being done because I feel like when we were writing like that's like for example, in my playwriting class, like this is art that people have.

They want to be creative about and they want to share to their class, they want to feel proud of it, um, so it could be connected to maybe issues that that they you know feel uh you know that they care a lot about or a personal experiences, and it would just be it would feel so hollow to, for example, use AI for that.

Or my classes, you know, we have these class discussions a lot.

My school's a consortium school, so that's probably why, right?

If you're using AI that to write something, you know, I don't think you're gonna be having a very compelling, let's say, defense of your opinions or your ideas, and you're not probably gonna be able to engage very well in the conversation.

It's gonna be boring.

And I think students like being able to share their views on various issues and back it up with their with their own ideas and with their own evidence.

Okay, thank you so much to this panel.

I appreciate your testimony today.

Um, are there are there any other students in the room that are that want to testify?

And you fill that a slip?

Okay, so this panel is dismissed, you can go back to your seats.

I want to thank you.

Um, you guys could come up.

I'm just gonna need to know what you're new.

Yeah.

Students come on up.

No, it almost reminds me of um like the AI technology is pretty tempting.

It's like very it almost reminds me of social media.

Like, like students, they want to like be on social media, but they uh they're I I spoke to a lot of students who are so happy when they banned the phones because it takes it, it takes it away from that because they want to be engaged with their friends.

Um, so are you are you also a student?

Are you testifying?

In the back row?

No?

Okay, just you.

Okay.

You're the only student uh okay.

Please state your name for the record, and you may begin your testimony.

Good afternoon.

My name is Vinny Dahl, and I'm 15 and I a sophomore at Queen's High School for the Sciences.

Today I come here to talk about two things.

Well, AI is part of it, but first I want to talk about introduction 684-2026 about geographic location and how it affects students.

First of all, education and the quality of education is not the only thing that affects students.

Depending on how far they live away from schools, it affects how well or how much education they get.

If they cannot come to school at time because of the MTA, they cannot receive a quality education.

These are the omnics be uh provided.

When these fail, students cannot receive uh equitable education.

Most students, especially with those with disabilities or special education, often live further away from their schools or often receive more struggles in getting to their schools.

Now, this there's this inequality can lead to a plateau of issues.

Now, I want to serve the rest of my time to talk about AI.

Artificial intelligence is not as easy or as simple as we are portraying it as.

Yes, thank you.

That was all him.

I don't know if he knows.

He's building it with AI.

Thank you so much for I know he's not, I'm kidding.

Um, thank you so much for your testimony today.

Thank you.

This one.

And as always, if I mispronounce your name, I apologize in advance.

Uh, you will correct me for the record.

Um, Tanaya Nasser Frederick, Edgar Edgar Trout, Travis Prox.

Oh, okay.

Okay.

Okay.

Hi, my name's uh Tanaya Nasser Frederick.

I'm uh with the Collaborative Research Center for Resilience, which conducts investigative and community-centered research to ensure the government use of technology advances the public good and democracy.

I'm also speaking as a member of the AI moratorium coalition.

The data privacy officer mentioned that schools have the right to share student data with school officials, but he neglected to mention that most vendors in New York City public schools are designated as school officials.

Uh the Department of Education does list its FERPA exam vendors on its website.

This list includes the mandated curriculum HMH, which hosts a mirror.

Amira has taken a major investment from the investor branch of Amazon's Alexa.

HMH is owned by the private equity firm Veritas Capital, which also currently owns the company Pariton that is contracted to build DHS's biometric database.

Um two companies consulted on the DOE's preliminary AI guide guidelines, etc.

and Salesforce also maintain quiet contracts with Customs and Border Patrol.

So the data privacy officer needs to better explain why data privacy agreements are not available to the public because it's a great cause of alarm for parents, teachers, and students who are concerned that student data is being used for third-party product development and even contributing to higher resolution databases used by the Department of Homeland Security.

The public also doesn't have a way of verifying whether a vendor has undergone an IRMA approval process.

We don't know whether HMH's AI add-ons writable in a mirror are IRMA approved, or if they are considered approved based on HMH's pre-existing approval.

And take a company like iReady that's facing multiple class action lawsuits around the country for wrongly sharing student data.

Even were there an overwhelming uh community mobilization against a vendor such as iReady, most districts are too dependent on the product to cancel their contract with the vendor.

The real concern around AI is private equity and venture capital back consolidations around private solutions for public education beyond siphoning public spending for the profit of these companies and their vendors.

A district's dependence on a vendor is a threat to democracy and to the kind of community engagement that the DOE claims to value once the district establishes a relationship to dependency with the vendor, the vendor is operating with impunity.

Thank you so much.

Good afternoon, chairs, committee staff, and the wonderful students in the room.

I would like you to consider four activity categories when debating a moratorium on AI.

Category one, the top-down rollout of ed tech tools by the central administration.

A pause is prudent here, especially given the insightful questions of this council.

The next three categories, though, should live beyond a moratorium.

Category two is the bottom-up learning innovations by classroom educators.

Imagine a lesson where students tackle an issue in their community, a gen AI conversation helps them aggregate public data and stakeholder views for them to discuss, analyze, and critically debate.

Category three is learning about AI itself.

Imagine a Gen AI tool writes a story, print it out on paper, view it screen free.

Students then learn to uncover any biases in that output.

Today's students must experience this as future tech leaders, policy advocates, and consumers who will vote with their dollars on technologies.

Category four, teachers using AI tools to support lesson development.

Imagine a teacher who prepares a supportive lesson after conversing with a Gen AI tool about students' common misperceptions on the topic.

Students benefit without FERPR-protected information being shared.

Encourage and support teachers to propose innovations and the AI tools they need.

Rigorously review the ideas from compatibility with regulations.

Accompany this with AI-specific professional development.

Strengthen it by organizing co-creation communities of teacher innovators, community stakeholders, and pedagogical innovators at higher education partners.

I can't stress the higher education portion enough.

As we think about what activities live outside a moratorium, we should remember that these technologies will continue to be used in the world.

Our students will graduate, they will become students in college where they'll confront AI.

We must be responsible in our use of AI, but we must also not do a disservice to our students in pausing the beneficial uses.

Thank you.

Thank you.

Travis Proof from Fordham University.

Uh, Chair De La Rosa, proud uh distinguished alum of Fordham University, good to be here today.

Um, thank you for the opportunity to talk.

You know, this hearing's tackling some of the critical operational questions the city is going to be having to grapple with, but our testimony at Fordham is focused on a bit more specific question, the number of the students very ably raised, which is when AI is in use, how can we ensure it is applied ethically responsibly and with end users trained to use it with humanity discernment and emotional intelligence?

We can teach students how to ask better questions, not simply generate better answers.

Um it is important to note the conversation we're having about the opportunities, challenges, and concerns of AI is one machine can technology cannot replicate.

That is both our power and is our blueprint.

There is a global reckoning happening around AI, and just weeks ago, Pope Leo proclaimed that the dignity of the person must always be present in the technology.

It's not by accident that the dignity of the person has not always been present.

Recent years, careers in social work, teaching, ministry, the arts have frequently been deprioritized.

I mean it's the push for the youth and young adults to pursue opportunities and STEM.

Fields understand understandably deemed a national priority, driven the need for economic competitiveness.

But in some corners, the humanities were outright stigmatized.

The lack of balance is why we need this important conversation now.

And at Fordham, we are following four general pillars of how should we approach AI in education.

One is understanding the foundations of AI, two, discerning ethical versus unethical content and uses.

Three, how to augment, not replace work product and workers.

And four, recognizing when the use of AI is not appropriate.

We're operationalizing these values through the construction and launch of an ethical AI hub at our Lincoln Center campus.

And I've included in my testimony a few of our key initial areas of focus, as well as some of the community engagement we've done over the past year and a half to help guide that work, and we'd be happy to share that information with you if it's helpful.

But the gist is this: we need a citywide training strategy that goes beyond not just education but the entire citywide community.

We want to be a partner.

Thank you.

Thank you so much.

Um I do want to highlight step three, Dr.

Trouton.

Your um your I guess your four recommendations.

And I think if you could talk a little more, just a little more about it, uh, because I think there's an assumption that, oh, to learn about AI, you need to put a chat bot in front of a kid, but you're you're highlighting that you don't need to use AI itself on a computer to learn about AI.

Is that correct?

One of the tenets of computer science education, which I'm intimately involved in, is what we call unplugged activities.

Just because you are learning about a computer does not mean you need to be sitting in front of the screen.

And I again stress that if you if we work with higher education partners, we can co-design these experiences that teach about AI without doing some of the damage of extra screen time.

Sorry, you were trying to figure out your name, but I couldn't catch it.

What's your name?

Tanaya.

Yeah.

Yeah.

No, I'm sorry.

Sorry to uh.

Yeah.

Um my question is uh you mentioned that the basically what you're describing to me sounds like a loophole in um in the vendor system where if they're registered in-house.

Can you explain that a little bit more?

Yeah, I mean, that's been going on since the 70s, like ever since FERPA was invented, they never really defined what a school official is.

So, you know, like the construction company that who they hire to do the cafeteria stuff, those are all school officials through like no child left behind and the every student succeeds act.

It became more formalized that like vendors are designated as school officials, and most ed tech vendors and classroom application solutions, all that stuff.

Uh they function as school officials.

And uh you can see that on the DOE's website.

But yeah, that's what the Yeah, that's what the data privacy officers didn't mention.

And I think that's an important detail to how uh data is shared to third parties, how schools can share data to companies without uh parental consent.

Okay, I just wanted to clarify that because it I thought I understood it correctly, but you explained it way clearer now.

Thank you.

I just just for at form, I just want to clarify similar to the question for Dr.

Chout, which is a citywide training strategy.

Is your contention that we should be training kindergarten?

You said K-12 students, teachers, administrators.

Um, is your contention that we should be training kindergarteners using a computer or on AI, or is there some other clarification point about what you mean by K-12 students empower them as creators, critical thinkers?

For our university, the stance is this we are not trying to convince people to use AI.

We are trying to convince people how to do that critical thinking that the students were talking about, you know, how to understand ethical versus unethical uses, whether that's on a computer, whether that's other technologies.

But that's really what begins for us, and I think that's all age groups, it's all sectors, it's outside of education, it's in within the small business, medium-sized enterprise space as well.

But that's really our position.

We're not here to convince people to use AI, we're here to convince people to understand ethical realities, discernment, and to understand when it's being used on them and if they do use it, how to use it well.

Thank you.

Thank you to this panel.

Thank you so much to this panel.

Appreciate your testimony and look forward to continued work together.

This is very clear that we need to do a lot of work and putting those guardrails in both academically and for safety purposes.

Our next panel, Ruth Loencron, Gracie Messier, Matt Henning, Larry Lieberman, and Liz Beth Lucero.

Ruth isn't here, just so you know, she's from my organization, so you may begin one already.

Please take your name for the record.

Good afternoon.

Thank you so much for the opportunity to submit this public comment.

My name is Lizbeth Lucero.

I'm the Deputy Director of Advancement and Communications at the Urban Assembly.

And I'm actually here on behalf of our Chief Executive Officer David Adams, who is delivering a commencement address at the Urban Assembly New York Harbor School.

And we really appreciate the council's focus on AI, student data, and privacy in our public schools.

At the Urban Assembly, we believe that AI has the potential to expand opportunities for students and educators, particularly in areas of college and career advising and instructional support.

At the same time, we agree that AI must be implemented thoughtfully with strong safeguards to protect our students and ensure responsible use.

From our experience, one simple lesson is true.

AI works best when implemented and it is human-guided.

These tools can help educators and students access information more effectively, but they cannot replace the relationships and expertise about counselors, teachers, and school leaders.

We also believe that student privacy is non-negotiable and tools reflect this, and our tools reflect these philosophies, which is why our schools have been recognized with EdSafe AI alliance for demonstrating responsible and ethical AI practices.

These safeguards ensure that innovation and accountability go hand in hand.

At the same time, a focus on risk should not prevent us from pursuing opportunities.

Our work with AI supported instructional feedback has led to measurable improvements in teacher quality, increasing the use of respectful language in the classroom, and an increase in lesson previews.

We've also seen increases in student participation through 14% of increase in student talk time, alongside a 14% decrease in classroom silence, which indicates an engaging and inclusive classroom environment.

Besides these numbers, our impact with students and teachers are real.

As AI becomes more common, the question is not about whether AI could be used for good and bad.

It's about how we can ensure to solve the biggest problems for our students.

So thank you again for helping us to figure out solutions and innovative strategies strategies for our public education system.

Thank you.

Thank you so much.

Good afternoon.

Please make sure your microphones are on.

Remember, red means go.

Yeah, one more good afternoon.

Yep, there you go.

Good afternoon.

There we go.

Good afternoon, and thank you, Council members.

My name is Larry Lieberman, and I'm here on behalf of Mouse, a nonprofit that for 30 years has trained New York City public school students and teachers to use technology creatively and responsibly.

We appreciate the council's attention to security and to artificial intelligence in the classroom.

The answer, of course, is not avoiding AI education or distancing our educators from positive and helpful technology.

The answer is more training.

Teachers and our young people are know what AI, know that AI is available and need to learn what it is, how it works, where they are using it, and how to use it safely, intentionally and equitably.

That is exactly what Mouse's AI League does at 40 New York City public schools across all five boroughs with curriculum train created in partnership with Carnegie Mellon University.

That, to your point earlier, includes an incredible set of generative AI lessons using only whiteboards, markers, and yellow stickies.

In the program, like Mouse's AI League, students in NYC are learning how AI works and knowing this helps protect them.

We urge the council to vigorously move to protect student safety while supporting high-quality digital learning.

Together, our jobs include ensuring that New York City public school students are prepared to be the future engineers and innovative innovators at New York City's formidable employees.

Last week, 83 New York City teachers in Council District 35 attended a mouse generative AI fluency training.

Yesterday, 25 more high school teachers in Council District 15 attended mouse generative AI training.

The demand among teachers for AI training is as large as anything we have ever seen in 30 years.

Please come visit Mouse in the Classroom.

Get to know our teachers.

Thank you.

Happy to Good afternoon.

My name is Gracie Messier, and I'm a legal intern with the Disability Justice Program at New York Lawyers for the Public Interest, or NOPI.

NOPE is here to testify and supports the goals of intro number 853 and appreciates the council's commitment to special education services.

However, we encourage the council to consider whether the study as currently proposed is the most effective use of this of the city's resources.

Much of the information the study seeks to collect is already available through existing reporting.

We encourage the council to refocus the study to answer questions that existing data cannot or devote funds directly towards improved special education services.

According to the city's most recent data, only 28% of English learner English language learned sorry, English language learners with individualized education plans or IEPs receive those services in another language.

And 95% of English speaking students were fully receiving their recommended special education services, while only 47% of Spanish speaking students were fully receiving those same services.

Data also show substantial differences in accessing services based on income.

Students receiving dis services disproportionately come from low-income households, and students from higher income neighborhoods are more likely to be placed in inclusive settings.

Gaps in services fall hardest on low-income families and multilingual households.

In NOPE's work, these communities have consistently reported the same barriers, delays in receiving related services, shortages of qualified providers, difficulty accessing bilingual services, and limited availability of appropriate placements.

Nopey shares the council's commitment to improving outcomes for students with disabilities.

However, existing data already established that language, income, and geography affect access to services.

Whether through investment, through direct investment, or more targeted study, we urge the council to instead focus its efforts on ensuring that students receive the services to which they're entitled.

Thank you for the opportunity to testify.

Thank you so much.

Good afternoon, Chair De La Rosa and Chair Denowitz.

My name is Matt Henning, and I'm the director of government affairs at Tech NYC.

Tech NYC represents hundreds of companies from global tech leaders to startups that are deeply committed to the advancements of the tech sector in New York City and supporting the city's public schools.

I'm pleased to testify today on pre-intro 2129 and 2130.

Tech NYC does not believe New York City's children should be treated as test subjects, and new technology should not be deployed without oversight, clear policies, or transparency.

But we are equally concerned about the risk of failing to prepare the next generation for a world where AI tools are pervasive.

Without thoughtful guidance from trusted institutions, students will engage with AI entirely on their own terms, not unlike what happened with social media.

The answer is not to choose between adoption and a prohibition, but to build a transparent, accountable framework that earns the confidence of families and educators.

That is precisely why we support the intent and goal of both of these bills.

However, Tech NYC does have a few recommendations to further improve these bills.

First, the definition of algorithmic tool and two one two nine is broad, and we'd encourage the council to consider whether additional clarity would help uh DOE and vendors focus on compliance resources and tools that genuinely implicate the concerns motivating these bills.

Second, these bills layer new requirements on top of existing Office of Algorithmic Accountability and IRMA framework.

Implementation guidance should make clear how they how they coordinate to avoid duplication.

Last on 2130, uh the first quarterly brief reporting deadline of October 31st is a fast moving timeline, and we encourage the council to work with DOE to ensure the necessary infrastructure is in place.

Thank you.

Thank you for your testimony, and thank you.

These are my two bills to increase transparency and accountability.

So I thank you for your recommendations to uh improve the bill.

Um, Lizbeth, I I'm I just want to understand something.

You're you're you support our our schools, and you've chosen um to uh use the ed safe framework.

So to me, tell me this is true or not.

To me, that means that DOE has not provided a robust enough support system.

So you've chosen to go one step further and and use like an outside um uh assessment tool.

Can you talk a little more about this decision?

Yeah, definitely.

So, our one of our uh AI tools is used for instructional quality, and it's been recognized by the EdCep AI Alliance.

Um, I'm not 100% sure the relationship that we currently have with them.

I uh I most of my notes are from our team who builds the tool um at the urban assembly, and um there have been conversations with um the NYC uh DOE and OSP with some of the work that we're doing in our ed tech space.

Um, so hopefully that clarifies some of the no it sounds to me like it's another evaluative uh uh tool, but we will look more into that.

Um thank you to this panel.

I want to thank this panel for testifying today, and we'll move on to our next panel.

18 Chu.

Someone with beautiful handwriting, Michelle Blonde.

Tell me the rest, please.

Blondville.

I say I love it.

It's in beautiful cursive.

I have to.

Blonde Monville.

Beautiful cursive.

I meant it when I said, yeah.

You know, they don't teach cursive in school anymore.

Catholic school.

Catholic school.

I maybe I'll send my kids to Catholic school.

Learn.

I guess.

No, that learns cursive, it'll be good.

Shannon Richie and Julie Zhang.

Okay, you may begin when ready.

Good afternoon, Chair Dinowitz and Chair De La Rosa.

Thank you for the opportunity to give my testimony.

My name is Yatin Shu, and I'm speaking today on behalf of Place MYC, an organization that advocates for high quality accelerated learning opportunities for students in New York City public schools.

Place NYC strongly supports intro 684, which calls for a study on how students' language status and geographic location affect access to early elementary gift and talented programs.

GNT programs are unevenly distributed across the city.

Many districts, particularly in the Bronx, Brooklyn, and parts of Queens have few GNT seats.

Geographic location should never determine whether a child receives an education they need.

Every year, 10,000 families apply hoping for a spot with just 2500 seats across the city.

There are simply too few seats.

Equally troubling is the sole reliance on a subjective teacher recommendation for GNT identification for the program's lottery.

This process is prone to teacher bias, and we believe it disproportionately harms immigrant families where students' home language is not English.

The current pre-K teacher assessment takes place just three months after students in the classroom.

As a result, many bright English language learner students may not be recommended for gifted and talented programs, even when they are clearly ready for more challenging work.

We urge the council to examine these problems and recommend concrete solutions, specifically the adoption of an objective evaluation measure in early grades.

Universal screening uses using standardized assessments will give every child a fair opportunity to be recognized regardless of language background or zip code.

By including an objective evaluation and early identification, we can ensure that students who are ready for acceleration receives the appropriate education they deserve.

Place NYC urges the council to advance this bill so we can finally gather the data needed to build a fairer and more effective GNT program for New York City students.

Thank you.

Thank you so much.

Michelle.

Yes, my name is Michelle Ann Blumenville.

I'm a health and fitness educator for 40 years, a former adjunct lecturer at New York University, former instructor at Bash Institute of Technology and at the American Red Cross.

I'm here today to represent those innocent American citizens who are being experimented on the use of AI and to add to the supercomputer.

Thank you for your servitude in these difficult times.

I'm speaking on behalf of everyday people who are Havana syndrome or anomalous health incident victims, some knowingly and unknowingly, with the glaring awareness of the benefits afforded to our diplomatic counterparts helping American victims affected by neurological attacks.

We certainly hope that's to the notion that one day we will be free from torture, pain, invisibility, and the weaponization of technology.

I do thank the New York City Council for all of those laws put into place, like the POST Act, Hanifa's law that she just put into place about uh businesses accessing our biometrics and so forth.

Everyday people, Havana syndrome victims is compromised of diagnosed Havana syndrome, public citizens who have been unlawfully experimented on and who endure targeting in various infamous matters.

These heinous crimes include but are not limited to organized stalking.

Thank you for those harassment laws also that are put into our workplace, smear campaigns, noise harassment, electronic assault from directed energy weapons, nonconsensual human experimentation socially and technologically as V2K blue eye technology and AI.

Thank you.

Thank you so much for your testimony.

Thank you.

Hello, my name is Shannon Richie.

I am a parent of a third grader in Greenpoint, Brooklyn, and a member of the SLT.

I'm here today to speak out against the use of AI in our schools and to encourage a push from the council to reduce our usage of ed tech products.

First, I would like to thank the council members who have signed on to our moratorium proposal.

Thank you for listening to your constituents.

To those who haven't yet signed, I hope you are listening today.

The DOE claims to want parent and student input, but they couldn't be bothered to stay here and hear us out.

So much of what they said today is misleading and misrepresenting what we are experiencing on a daily basis.

It's nearly impossible to find real peer review data supporting the use of AI in education.

So what are we doing here?

Why aren't we listening to scientists, researchers, or education experts?

Why are we simply trusting the words of salespeople of representatives from ed tech and AI companies?

According to a New York Times article, Google claimed, quote, investing in schools held onboard kids into Google's ecosystem, end quote.

Schools are force-feeding kids tech products, and those products are collecting, storing, and selling private information pulled from our children's brains.

This is a plot straight from a dystopian sci-fi novel, yet we are living this reality every day when we send our children to school.

Our kids and our teachers deserve more than being reduced to data collection.

Speaking as an educator myself, learning is driven by human connection.

It is not merely pouring information from one source into another, which is essentially how ed tech products work.

If we want a robust public education system, we need to nurture it and curb the overreliance of on screens.

We need an AI moratorium now, and the city needs to stop seeing our children as products being ripened for harvest by Google and predatory ed tech companies.

Thank you so much.

I have to ask first.

I have to ask, you said you're an educator yourself.

I'm not currently teaching, but I was a teacher, yes.

Very cool.

Okay.

Thank you.

You may be in, yeah.

Good afternoon, Chair.

Thank you.

Good afternoon, Chair Dino with the members of the educational committees.

My name is Judy Zen.

I am a program assistant at Homecraft Community Services, Montisocial Services Center.

I am also a mother of two daughters in New York City Public Schools.

Thank you for the opportunity to testify in support of Intel 684 and Intel 853.

The education for all bills introduced by council members Suzanne Zuang.

Homecraft community services serve low-income immigrants and limited English professional families across Southern Brooklyn.

Many parents in our community care deeply about their children's educations, but they do not always have equal access to clear information.

Into 853, we require the city to study how language needs income and locational effect access to special education programs and services.

A child's opportunity should not depend on whether a parent speaks English fully, understands the school system, or lives in a neighborhood with more resources.

In our borough neighborhoods, in Brooklyn, families face many barriers.

This includes namely outreach, legal translated materials, transportation challenges, and uneven program availabilities.

Thank you so much.

Thank you for your testimony.

And as I said, you can submit the rest in writing.

Okay, thank you so much for your testimony.

I like to thank this panel for their testimony.

Uh this evening.

And our next panel.

Kaiser, Laney Hameson, Kelly Clancy, and Liat Olenic.

Yeah.

Okay, so the So Kaiser and Laney will testify.

Thank you.

Go ahead.

Hi, my name's Lainey Hameson.

I'm the co-chair of the Parent Coalition for Student Privacy.

Thank you for holding these hearings today.

I wanted to start by addressing some of the misinformation that was spread by DOE this morning.

Google Gemini is used in hundreds of elementary school classrooms, installed on thousands of young children's Chromebooks and inserted into their education every day.

So I have no idea what the DOE official meant when he said it was only used in a few high schools.

Google Gemini was also cited just this week as one of the most unsafe chat bots on the market when it comes to mental health, etc.

Um, I also wanted to deal talk about the Power School and Illuminate breaches, which release the data of many, many thousands of former students.

That data should no longer have been held by the vendor.

The state law requires data minimization and data deletion when that data is no longer needed to carry out services, and yet the DOE does not require this.

Power School has been noted over and over of not um complying with the most basic security protections, which led to the breaches of uh millions of students' personal data a few years ago, and yet the DOE allows um schools to still use more than 15 power school products, including novience, which more recently was cited as disclosing data illegally in violation of state laws and federal wiretapping laws, and a 17.25 million dollar settlement was made by the court, which New York City families are eligible for, and yet the DOE has not told any families about this alleged breach or that they're eligible for a piece of the settlement.

Um, I just wanted to go on a little bit more.

My my testimony is very lengthy, but the DOE has not been following compliant with the privacy laws.

That's why we recommended strengthening them, and they did not strengthen, they did not use the recommendations that we proposed on the AI working group.

Thank you very much.

Kaiser.

Good afternoon.

My name is Kaiser, community organizer with the Alliance for Quality Education.

AQE supports intro 853.

We oppose intro 684.

All academic programs and services should be equitably accessible, but there is no evidence that children eight and under can be reliably identified as gifted, or that separating them from their peers is educationally sound.

AQE encourages council to strengthen and pass the preconsidered bills to propose annual reporting and quarterly data breach reporting.

There is no data showing that AI learning tools are effective or safe, except claims provided by the profiteers.

Yet the Chancellor has announced plans to quote beta test these products on our children, including in District 75.

The bills should include public disclosure of evidence of efficacy.

Algorithmic tools have been proven to have racial and gender bias.

The bills must require assessments of bias and hallucinations in all AI tools in use.

Our kids' best interests, not profit, must be centered here.

The bill should include public disclosure of the number and positions of all DOE staff who have existing financial or other relationships with the companies that profit off of these products, and those companies' involvement in developing trainings and policy.

Unfortunately, reporting is not enough, especially when we can't trust it.

There were many inconsistencies in DOE testimony today.

Flagging staff either not having thorough understanding of what their department is doing or misrepresenting it entirely.

AQE is grateful that more than half of the council has signed on to the call for a two-year moratorium for all AI in schools.

Right now, kids are losing access to the most essential kinds of learning backed by extensive research, open-ended play, hands-on inquiry, and culturally responsive teaching, differentiated by skilled human teachers to suit their specific population of students and their layered human stories.

Parents don't know what companies have their kids' data or products are being used directly by their children, used to surveil them or used to shape their instruction.

And when they ask, they're told DOE either doesn't know or can't say due to NDAs.

Thank you so much.

These unproven biased and dangerous products are sucking up valuable classroom time and public funding shed a light on some of the issues going on with the DOE, your testimony today.

And as I mentioned at the round table.

So I want to thank you.

I want to thank you both for your testimony.

And remind you again, written testimony can be submitted to the council at testimony at council.nyc.gov.

Thank you so much.

Our next panel.

Stephen Stowe.

Beckham Cowley, Lisa Walker.

What is one?

Jay Editon and Molly Sinak.

I want I want to mix things up.

Let's start on my right this time.

Alright, go ahead, Molly.

Great.

Thank you so much, Chair.

Chairs.

Good evening.

My name is Molly Senek.

I am the education and employment community organizer at Center for Independence of the Disabled New York.

Advancement in digital technology have provided critical access for people with disabilities to spaces that were previously far less accessible.

It has meant the ability to participate in activities remotely, more consistent use of assistive technology, and the immediate availability of practically infinite information, all things that can be and are incredibly valuable in an education setting.

However, digital technology has also made it easier to harass people and to perpetuate bias and to steal personal information, to spread or even generate misinformation and to erode many of the skills like critical thinking that schools are supposed to build.

Over the last few years, New York has struggled to contend with having to retroactively set up guardrails inside the classroom for technology that has already had the chance to become ubiquitous outside the classroom, which obviously we know includes a cell phone ban and an attempt to restrict social media usage for children and teens.

Now there is a new technology that is developing rapidly enough that it's causing decision makers everywhere to rush to incorporate its usage for fear of being left behind without necessarily considering that we actually have an opportunity to get ahead and avoid the problems that we saw with cell phones and with social media.

As of right now, AI is developing so rapidly that we do not know enough about it or about the impact of its long-term usage on students, especially students with disabilities, to begin incorporating it into our schools without doing our due diligence.

While we cannot necessarily protect students outside the classroom, it remains our responsibility to protect them as best we can within it.

Just a couple of things, more details are in my written testimony.

We're looking at the impact of critical thinking, bias, misinformation, mental health, trust erosion, and also the conflation of assistive technology with AI, meaning that we don't yet know whether it masks critical symptoms and potentially delays diagnoses.

Thank you.

Thank you.

Hi, my name is Lisa Walker, and I'm a mom.

So earlier, the first deputy chancellor said that YouTube is not on DOE devices.

YouTube is on DOE devices.

Both my seventh grader and first grader have DOE-issued devices.

And as far as I understand, according to New York State law, I thought YouTube is not supposed to be on these devices.

But at my school, every child has a device sent home with them as young as four years old on weekends.

I can see on these DOE devices that my children have full access to the internet, YouTube, and no real safeguards.

Parents can opt out of a device being sent home, but it's still there waiting for them at school.

The DOE says, well, at least they're software installed, so they can't watch porn.

Well, that's just not good enough.

The DOE has not fulfilled their promise to keep kids safe and to protect kids' privacy.

We use two apps at our school against our consent.

We use two apps that are s at our school.

One is AI powered Amira, a cute avatar that talks to children collecting their voices.

Many parents don't even know their kids are using ed tech apps at school or using AI and being recorded.

Instead of putting the onus on parents to ask questions, parents deserve transparency and should be told what technology their kids are interacting with at school.

And we should also be asked for our consent.

Our kids deserve an education focused on real social connection to teachers and students, not AI.

Thank you.

Thank you very much.

Okay, uh, thank you.

And thank you, by the way, for getting my name right.

That's a rare and exciting occurrence.

My name is Jay Eddon.

I'm a freelance creative professional and the parent of a New York City student.

And I'm speaking here today to urge you to adopt a moratorium on the use of predatory LLMs, better known as AI in New York City schools.

So the case against AI schools is really substantial, so much that I can't fit into two minutes, so I'm gonna sort of focus down on something that people haven't brought up so far.

Specifically that the largest large language models are built using the stolen work of thousands upon thousands of writers and artists.

And especially in a city where so much of so many of us make our homes, the adoption of technology built on our non-consensual, uncompensated labor and intellectual property is at best an insult.

But there's a bigger problem attached to it, which is the fact that adopting that technology into schools means trusting those same predatory companies who stole all that information and all that work, unapologetically, with our children's sensitive data.

We know that AI also undercuts children's social and emotional development, and I want to talk to in particular about its impact on adolescents, because they're the age group targeted commercially and in our schools with the highest concentration of AI.

And it's it's bad.

It's that's the point that the APA has issued a health warning about adolescents and AI, and that in itself isn't much of a surprise given that we've seen documentation in the last year that chatbots are capable of inducing psychosis in previously healthy adults and have driven multiple children to suicide.

My kid um isn't an adolescent, so she's just about to start pre-K, where tech magnets are already pushing and increasing use of AI.

As you mentioned, kids as young as four are getting devices to take home.

Um, and as a parent, I want my kid to actually learn at school, not just facts, but to think and reason for herself to solve problems and to grow as a human being.

Um, and introducing AI as it now exists in New York City classrooms will stunt that same critical growth for my kid and for thousands of others.

Thanks for your time.

Thank you so much for your testimony.

And you may begin.

Thank you, Councilmember Dinowitz and uh Dila Rosa and um other members of the Education and Technology Committee.

My name is Steven Stowe.

I'm a parent of two former public school students, and I previously served on a community education council.

I'm here to support bills INT 684 and INT 853, sponsored by Councilmember Zhuang.

I agree with the premise of both these bills that many um issues in our education system come simply from a lack of awareness and knowledge of what's available to parents.

It's very concerning that white families get the majority of private school special education funding, as Chalkbeat reported.

Went to white families, despite the fact that they make up only 12.5% of students with disabilities.

By comparison, as Councilwoman Zhong was highlighting earlier, black and brown students make up 75% of students with disabilities, but only got 24% of the tuition.

This is due to very complex cultural, socioeconomic reasons and educational reasons, in short, an awareness gap exists.

The City Council can take the lead on addressing these inequalities by advancing INT 853 on the GNT.

For years, GNT programs were systematically cut in black and brown neighborhoods for political reasons.

Perhaps unsurprisingly, we also got a steady reduction in the number of black and brown students attending the most selective high schools.

I applaud the prior administration for starting to expand the number of programs in underserved districts, but more should be done.

There's still not enough awareness of these programs, and we can't just make token efforts.

Continue efforts to address gaps and awareness, which this bill will help do.

The council can take the lead on addressing this inequality by advancing INT 684.

In conclusion, the DOE does offer an amazing range of programs.

So much of the issue is knowledge and awareness with parents.

I support these bills and hope they are passed, acted on, and used to improve real education policy in New York City.

Thank you.

Thank you all for your testimony.

I will now call the next panel.

Greg Pack.

Kristen.

Adrian.

Savore.

Carrie Hill.

Ariana Rebellini, and Effie Zachary.

Oh, I do.

You may begin when you're ready.

Okay.

Thank you.

Chair Dinowitz and Chair Del Rosa.

My name is Effie Zachary.

Formerly I used to be the vice president of CCHS, and more recently, the Queen's representative on the PEP.

I was elected by parents to both positions, not appointed by any politician.

I wasn't going to speak, but I heard a lot today that made me do want to speak a little bit.

Almost everything we heard today about contracts and privacy, nothing is new.

It was discussed at least three, four years ago, especially when the Illuminate had a serious breach.

I was on many committees, many task force.

We had questions, we gave a lot of suggestions, and I'm dismayed to see that almost every good sensible suggestion that was made has been pretty much ignored, certainly not implemented.

And to when I hear the term, like almost two year moratorium, it almost makes me laugh almost by default.

Everything that the DOE starts to do has a two-year moratorium that they're so slow to move.

Jokes are post with what I consider one of the slowest moving agencies in New York City.

Trying to contain something that moves very fast.

But we need to think like what are the guardrails that we need to put in place.

That should be the focus.

Thank you very much.

Thank you so much.

I appreciate you having us uh allowing us to speak.

My name is Greg Pach.

I am a comic book writer, uh, and also a New York City resident, uh, at a child go through the public schools, product of public schools myself.

I'm here because I believe that AI in the schools is child abuse.

It is a form of child abuse.

I don't use those terms lightly.

Writing is thinking, drawing is seeing.

If you get a kid to type in a prompt and generate text, type in a prompt and generate an image, and you say good, you have failed.

In fact, you have denied that child an exper a human experience of thinking and learning and doing stuff.

Kids are not products, and they're not in school to create products.

Kids are in school to learn, to go through the process.

We should be about process, not products.

Um, and uh AI is also a tool of liars, scammers, and fraud.

This is its primary use case.

And having kids use AI in schools sets them up to become losers who can't think for themselves and and and can't do things, and cannot be part of a community.

When a kid draws, a kid has to look at you.

A kid studies your face, a kid takes note of the world and the people around them.

If you deny a kid that chance by having them just type in a prompt, they learn nothing, and they do not participate in a wider community.

AI is child abuse.

I'll say it again.

Um, I also just wanted to take note.

So I support the AI moratorium in the public schools, and in fact, we should make it an uh uh, we should just we should just burn all those machines.

Um, I also want to take note of the fact that uh that these um, there's a racism button on these machines, and the billionaires can push that racism button on the machines as they have, and we don't want to thank you, and thank you.

No, you don't.

I think you are learning something from our president.

Calling people, please.

Thank you, Chairs.

Um, I'm Ariana Rebellini.

I'm a mother to a soon-to-be second grader in Queens and a writer, and I want to speak today about teachers.

My entire life I've had people tell me my father changed their lives.

Carlo taught high school English for 43 impassioned years, and if there's one thing I know beyond all doubt, it is that my father loved teaching and his students loved him.

When he retired, we set up a website where former students could leave well wishes, and it was a gift to see what they had each held on to.

From high-fives in the hallways to a new respect for reading.

My dad was born and raised in what he loves to call the boogie down Bronx, and rap has been a huge part of his life and syllabi.

Multiple posts shouted out Wu Tang Wednesday and Kendrick Friday.

And I say all of this to highlight the varied and immeasurable impact of teachers and the incalculable losses we in our community will endure if we declare them redundant in the face of AI.

If my father changed his students' lives, it was in ways that can't be replicated by a machine.

It is something living and indelible and unpredictable.

I can't believe I need to say this, but educators should be people.

Anyone who cares about children's well-being knows that.

I know you know that.

I know the DOE knows that.

This is a disingenuous conversation.

Last year I published a book about suicide, so I feel uniquely qualified to emphasize the dire emotional and material consequences of losing human connection.

Isolation and alienation are killing us, and I must note that black children are especially susceptible.

Chatbots, which we are already seeing being used as study guides and tutors are certainly not helping.

Our children cannot afford to lose even one person who is invested in their future.

I want to close with one message from my father's former student, Roshin.

She wrote, Your classroom is not just where I learned how to write essays or identify sentences, but it was where I learned how to be human.

So when we talk about how AI is destroying the humanities, we're describing not only academics, but more profoundly the very substance of life.

It is your responsibility to protect our children's right to both, and the moratorium will be a good start.

Thank you.

You say your dad's from the Bronx?

Yeah, he is.

Both parents.

Both parents.

Yeah.

Now you gotta tell me where in the Bronx.

Um Allerton Avenue was what they were near.

I was born in the Bronx, but you're part of the Bron.

You know, you're welcome back anytime.

Thank you.

All right, awesome.

Thank you.

Thank you to this panel.

Thank you.

Okay, I'm gonna call the next panel Iman God, Charlotte Pope, Andrea Ortiz, Anita Rad.

This side?

Okay.

Good afternoon.

I'll let you all decide.

Fielder's choice.

Thank you, Andrea.

Good afternoon, Chair Dinowitz, um, Chair de la Rosa and members of the committees.

My name is Iman Gadd, and I am the policy and advocacy manager at Girls for Gender Equity, a black feminist organization that works to reshape culture and policy through advocacy, youth-centered programming, and narrative shift.

So at a time when technology and AI are moving at an unprecedented pace, we are here today to urge the cities to prioritize the education, privacy, and safety of our students by redirecting all funds currently going to tech companies and surveillance tools towards organizations that are doing the work to support young people's tech literacy and towards proven safety infrastructure such as restorative justice, mental health access, and immigrant protections.

Last November, the Alliance on Gender Equity, our youth-led organizing base, held the town hall to inform the School Girls' Demand campaign.

A campaign with the purpose of uplifting the voices of young people and giving them the opportunity to ask their city for what they need.

AI and tech were hot topics in that room of 70 students from schools across the five boroughs.

Students expressed grievances that included the need for best practices for utilizing AI and learning environments, the racial disparities in AI access and education, climate-conscious ways to use AI, and the desire to learn about data privacy, along with their digital footprint and methods to protect themselves.

As young people from our youth programs reflected on their time and programs and the lessons they walked away with, one student said, I feel like one way that I can achieve social justice is by telling my friends and family about consentful tech.

I feel like that's very important to me now that AI is taking over.

Making sure that you're using apps and devices that ask for consent before they just take the information that you're giving them.

Keeping our youth tech literate and informed about the importance of data privacy was a high priority for this year's program curriculum to meet the moment and heed the call of our youth.

While small small organizations are doing the work to keep young people protected, the city needs to support these efforts by funding programming and safety infrastructures in schools.

This is why we call on the city to fund our reliance on gender equity and the amount of 300,000 to continue equipping young black girls, queer and gender expensive youth to fight for New York City that is resilient, and I'll put the rest of my RAM testimony.

Thank you so much for your testimony.

Hello, my name is Charlotte Pope.

I'm with the New York Civil Liberties Union.

Thank you for convening this hearing.

My comments focus on the student surveillance industry, where AI represents an imprecise blanket marketing term with no set meeting.

A note up top that we foiled for non-disclosure and data processing agreements, and they are full of redactions under the trade secrets exemption and the refusal of ed tech companies to explain how their AI systems were built and what they intend to do long term, protects these vendors from public scrutiny and has enabled big data systems that claim to do one thing to accomplish another thing, which is the accumulation of student information and public money.

Even where we can explain how a system works, there's still the question of whether certain systems should exist in schools at all.

Schools are testing grounds for AI products that monitor school issued devices, even simply typing and drafting of ideas, tracking physical movements and social connections and friend groups, and synthesizing moment-to-moment personal data spanning across years.

Some companies claim their AI surveillance systems can measure and track student emotions like aggression or generate an assigned mental health wellness force.

All of this produces floods of sensitive student data for private harvesting and opens the door to a digitalized school to prison pipeline.

It was confusing to hear today that DOE is not sure what surveillance software is in use other than Go Guardian, deferring that those purchases are made school by school, but also somehow very confident that certain GoGuardian products aren't in use and haven't been purchased.

As you asked, in the AI guidance, surveillance and behavioral monitoring products are both named as beyond the scope of the guidance, but also never allowed.

But what is their definition of behavioral monitoring and what is the mechanism for getting these prohibited products out?

When students complained about AI enhanced bathroom tracking apps, which is surveillance, DOE deferred to principal discretion.

The guidance claims in the DOE said again today that these apps must serve a clearly defined educational purpose, but we doubt that schools are convincingly articulating what they're asking these private surveillance sector companies to solve for.

Thank you.

Thank you so much for your testimony.

Hello, I'm Andrea from Dignity in Schools New York.

Public schools should be places where students feel safe and supported to learn, yet excessive policing and surveillance continue to undermine students' safety and well-being.

While we appreciate Mayor Mamdani's executive budget's proposed reductions to school policing, we are deeply concerned by proposed cuts to restorative justice, the mental health continuum, and immigrant family outreach.

It is a missed opportunity not to reinvest the savings from school policing into protecting and baselining critical services.

The executive budget also fails to address schools' continued reliance on metal detectors scanning and surveillance systems that create barriers to learning and disproportionately impact black and immigrant students.

Investigations have found that while 33% of high school students go through metal detectors, 48% are black high schoolers, which contribute to the fact that New York City black students that in New York City black students make up less than 20% of the enrollment but accounted for 45% of superintendent suspensions and 52% of police interventions.

Therefore, we are concerned by growing calls to expand school policing, scanning and surveillance in the name of safety, despite clear evidence of their disparate impact.

Schools are also increasingly adopting AI technologies that raise serious concerns about student privacy, data security, and surveillance.

Many AI-powered educational tools collect extensive information about students.

AI practices are particularly concerning for immigrant families at a time when federal immigration policies are expanding AI driven surveillance and data sharing systems.

Families have literat little transparency into how students' data is collected, stored, or shared, and these practices risk undermining student privacy protections under education law 2D.

Rather than investing in tech that monitors students, schools should invest in human relationships and support systems that helps students thrive.

We call on the city to stop buying student surveillance and scanning equipment and create a plan to use those funds to institute a student support staff ratio of no greater than 1100, freeze hiring for school safety agents and eliminate vacation uh positions, generating 90 million dollars in annual savings and using that savings to expand school-based restorative justice, mental health access, including 80 million to hire school climate coordinators in 500 schools and enact a two-year moratorium on the use of AI.

Thank you.

Thank you for your testimony.

Thank you.

Anita Rao.

Are you ready?

Okay.

Okay, you'll you'll go after the Zoom.

And Dr.

Kalidis Salas.

Good afternoon.

Thank you, Chair Dinowitz and Chair de la Rosa.

My name is Kalidi Salas Ramirez, pronounced she here ella.

I am gonna speak a little bit in the first minute about the work that the community education council in District 4 have been doing around providing feedback and policy recommendations for the Department of Education, not just around AI, which we have explicitly asked for a two-year moratorium.

In addition, we are also asking for a reduction in screen time in our schools, as well as an audit on iReady, which also has AI embedded, but we have found that it is an ineffective tool that we have spent over 100 million dollars in in contracts to be able to do screeners and assessments in our students that again are not informing us of proficiency and growth.

And so in turn, um, when thinking about again the investments that we make in our students and trying to understand what helps students thrive, these tools, quote unquote tools that the Department of Education claims to use to assess our kids, are ineffective, and that has been shown by the work of Dr.

Jared Horvath, who continues to advocate for the fact that ed tech and the implementation of ed tech into our schools in the last 10 years have significantly harmed our students.

I want to highlight also the work that, and so I'm gonna shift from CEC to Dr.

Galeri Salas, neuroscientist and developmental behavioral neuroscientists to say that now we know that adolescence has been extended until the age of 30.

The reason for that is is because we are implementing these computers and ed tech into our classrooms and not giving students the opportunity to engage in experiential learning and engage in the science of learning, which means that we need to take them away from screens and give them the experiences that they need and the relationships with their teachers and their community to develop the appropriate cognitive processes to be ready for an adult world.

And unfortunately, we're not doing that in our classrooms by not providing with them the appropriate resources to do that.

In addition to implementing a tool that is racially biased and impacting the diversity of our New York City public school students.

Thank you.

Thank you so much for your testimony.

We will now, that concludes the in-person portion of our public testimony.

We will now move to remote testimony.

If you are testifying remotely, please listen for your name to be called.

Once your name is called, a member of our staff will unmute you.

You may then start your testimony once the sergeant at arms sets the clock and cues you to begin.

Professor Julius Stoyanovich, you may begin once you are unmuted and the sergeant cues you may I start?

You may begin.

All right.

Thank you very much for the opportunity to testify.

Uh my name is Julius Dianovich.

I'm a professor of computer science and data science at NYU and the founding director of the Center for Responsible AI.

I have worked on AI governance in the city for nearly a decade, and I'm a proud New Yorker.

I have three asks today, and they share one principle.

The city should not run AI systems on our children that no one outside the vendor can see, check, or hold to account.

First, prove it works.

AI tutors are flooding into classrooms right now fastest into low-income schools because they promise to do more with less.

And today the vendors grade their own homework.

They self-attest that their tools are tested, equitable, and safe.

There is no independent yardstick.

The city's own vetting process, Irma, checks data privacy, but by its own account, it does not yet review whether a tool is effective or whether it's biased.

One in three New York City third graders cannot read proficiently.

A tutor that's quietly worse for multilingual learners scales that gap across thousands of kids at once.

So my first ask before a tool reaches a classroom, require it to meet an independent efficacy and equity standard.

Does it measurably improve learning?

And does it work as well for English language learners and students with disabilities?

Meets the standard should be a precondition for procurement.

Second, keep the data home.

Those same AI tutors record how children think, struggle, and learn.

Some of the most sensitive data we collect on a young person.

Today, that data routinely flows out through vendor APIs to models running on private platforms, where it can be retained and reused with little transparency and little recourse.

When I testified before the US Congress on Deep Seek, I call this a loss of data sovereignty.

The moment the data leaves your control and falls out of the time expired.

Okay.

The third is to keep the data open so that people can check whether these systems work.

So in summary, prove it works, keep the data home, and open the data.

Thank you so much for your testimony, and if there's anything to expand upon, please submit written testimony.

Thank you.

You may begin.

Hi folks.

Good afternoon, Chair Zenowitz and De La Rosa.

My name is Aisha Erfan.

I'm a proud product of New York City Public Schools, a former Gail Brewer and Google AI policy staffer, and the founder of Augmented Intelligence Advisory, where we partner with schools to protect student cognition.

New York City parents and electeds are organizing en masse against AI in schools out of a very real fear that untethered AI will cause students to lose the practice of thinking for themselves.

They're not wrong.

Recent studies by Microsoft, Carnegie Mellon, and Anthropic show that AI reliance reduces critical thinking effort and weakens conceptual understanding and experienced professionals.

If these are the effect on adults, the risk for students building foundational skills are even higher.

Most concerningly, untethered AI use threatens to create a cognitive divide along socioeconomic lines.

In tech heavy households, adults model guarded AI use, and others, AI replaces thinking entirely.

The answer is not a blanket ban, it's pedagogy.

I urge two immediate moves.

One, deploy cognition-focused AI literacy in classrooms across the city, where students you practice using AI tools under the guidance of teachers.

This means students forming their own hypotheses, interrogating what they know before reaching for a chatbot and creating AI output as a draft to push back against rather than a final answer.

Teachers are the experts in how learning happens, they know when a tool builds reasoning and when it short-circuits it.

Teachers, not vendors must guide this work in every classroom across the city.