The Pittsburgh City Council held a virtual-only post-agenda discussion on surveillance technologies on Monday, September 21, 2026, at 10:00 AM. Councilmember Deborah Gross chaired the meeting, which featured invited national experts: Dr. Lorrie Cranor from Carnegie Mellon University, Clarence Okoh from TechTonic Justice, and Councilmember Mike Siegel from Austin, TX. (Sarah Hamid from the Electronic Frontier Foundation was unable to attend due to illness.) The discussion aimed to inform Pittsburgh’s emerging surveillance technology policy, building on a September 2, 2026 report submitted by the city.
Dr. Lorrie Cranor (Carnegie Mellon University) gave an extensive overview of modern surveillance technologies—including AI-powered facial recognition, license plate readers (ALPRs), mobile phone tracking, smart home devices, and radio frequency sensing—and their privacy implications. She noted that AI face recognition is often inaccurate and can lead to wrongful arrests. She critiqued the City of Pittsburgh’s surveillance technology report for not systematically documenting what data is collected, how long it is retained, who has access, and whether access is logged and audited. She also identified a gap: the report does not cover independent authorities such as the Pittsburgh Parking Authority and Pittsburgh Water and Sewer Authority, which also collect sensitive data. She offered CMU’s privacy engineering program to assist with future analysis.
Clarence Okoh (TechTonic Justice) framed mass surveillance as a civil and human rights crisis, citing examples like a Maryland school district’s AI weapons detection system that misidentified a bag of Doritos as a handgun, leading to a traumatizing police response. He detailed the Pasco County (Florida) predictive policing program, which used secret data sharing agreements and an algorithm to assign risk scores to over 17,000 students (about one in five in the district). The program resulted in higher rates of exclusionary discipline and was designed to pressure families to move or sue—a practice later found to violate the Fourth Amendment. Okoh urged a comprehensive approach: prevent AI injustice by banning inherently harmful technologies, protect rights through testing and impact assessments, and repair harms through accountability mechanisms. He supported creating an AI Bill of Rights and community co-governance models.
Councilmember Mike Siegel (Austin, TX) described Austin’s experience ending its ALPR (Flock and Axon) contracts in June 2025 after discovering that Flock’s contract contained loopholes allowing data to be fed into AI systems despite a stated 7-day retention limit, and that 20% of police searches were non-compliant with city council guidelines. He noted that Austin’s surveillance impact report revealed 126 different technologies subject to their Trust Act. He emphasized that local governments are often not equipped to evaluate these tools and that even well-intentioned uses (e.g., park security cameras) can have unintended surveillance consequences. He advocated for transparency measures (surveillance impact reports) as an intermediate step for communities not ready to cancel contracts, and warned that data held by the city can be requested by state or federal authorities, especially given the current political climate.
Good morning and welcome to Pittsburgh City Council's virtual cable cast post agenda on surveillance technologies. My name is Councilwoman Deborah Gross, and I will be chairing this post-agenda discussion today. For the record, today we are currently joined by Councilman Kahari Mosley, Councilwoman Erica Strasberger, and we may have other council members who are watching or will be joining us later. And I would um also welcome our guests and appreciate you for making time today. We are virtual only because we have some national experts joining us from around the country. Um, and it was easier certainly to have this post agenda on a timely basis by having you stay where you are rather than trying to come and join us here in Pittsburgh. Um I am looking for our panelists. I'm going to um read a short bio for each of our panelists, um, and then I'll be handing it over to Professor Craner from Carnegie Millen University. Um first, I'd like to point out that we had a last-minute cancellation by Sarah Hamid, who is director of strategic campaigns at Electronic Frontier Foundation. We hope that she's feeling better soon. Um works with policymakers and communities to evaluate how emerging technologies affect privacy, which is one of the main concerns of this post agenda, um how they affect privacy, civil liberties, equity, and democratic governance. Um kind of intersect. Um, and AI. Councilman Siegel was elected in 2024. He has devoted 25 years to public service, including as a public school teacher, nonprofit leader, and subvoids lawyer. He led the effort to end the City of Austin's automatic license plate reader program and co-authored the Trust Act, which created mandatory transparency measures to evaluate the city's use of surveillance technologies. So again, we're pleased that all of our participants could join us. Professor Craner, would you mind kicking us off? I know that we were supposed to have Electronic Frontier Foundation before you, but we would be pleased if you would help start us off. Sure, thank you. I want to thank Councilwoman Gross for convening this important discussion and inviting me to participate. And I was asked to give an overview of various types of surveillance technologies that you should be aware of. And so I prepared some of that. And at the end, I will provide some brief thoughts on the Pittsburgh surveillance technology report. So it used to be that the main types of surveillance technology were cameras and microphones, sometimes quite visible, sometimes hidden, so that people would not know they were being secretly surveilled. Phone lines could be tapped, and people could wear hidden microphones under their clothes to secretly record conversations. All of this required some significant effort to carry out, and generally in the US, there were legal procedures that had to be followed. But today there are many technologies that are surveilling us all the time. We still have surveillance cameras in the form of security cameras on homes, businesses, government buildings, and public places. But now many homes are also equipped with smart doorbells that double as surveillance cameras and record not only people who come up to the front door, but also many people who happen to be passing by the home. It used to be that surveillance camera footage required a lot of time and effort for a person to review, and identifying people who appeared in the video was fairly difficult, especially in cases where unexpected people appeared. But AI-powered face recognition technology makes it increasingly easy to identify people from surveillance video and also to automatically label their actions. Of course, automated face recognition is not always accurate. In fact, it's often not accurate. And when this technology is used blindly for law enforcement, it can lead to wrongful arrests and other very serious consequences. Besides looking for faces, surveillance cameras can track vehicles by scanning license plate numbers. This can be quite useful for parking enforcement and tracking stolen vehicles. It's also, it can also assist with tracking criminal suspects. But this sort of data has also been used for stalking and harassment as well as for inferring all sorts of information about people, including travel patterns that may suggest medical treatments, extra matter extramarital affairs, or other highly personal information that people have a right to keep private. Most people today carry around mobile phones, which also serve as tracking devices. Wireless phone carriers track the cell towers that phones are near so they can deliver phone calls. The phones track their own location to provide location-based services to their users. However, this location data may be accessed and transmitted by mobile apps in some cases without the user's knowledge. In other cases, the user permits sharing of their location data, for example, with other users of a running or exercise app, but may still be surprised when that data is mined for completely unrelated purposes. Social media platforms may use location data to inform users about other nearby users on the platform, but also to discover people who are frequently co-located and then for relationships between them. Even when location data collection and sharing is done with user consent, there are often concerns about whether the user understood what they actually consented to. Privacy notices are notoriously confusing, full of jargon, very long, and consent interfaces are often designed in deceptive ways that nudge users to grant consent without actually understanding what they are consenting to. Newer cars also have built-in location trackers, which can be used for a variety of useful purposes, such as providing maps and locating stolen vehicles. However, they also risk turning the car into yet another surveillance device. When you connect your phone to a car, it may gain access to information on your phone as well.