The Surveillance Technology Surveillance Project, a New York privacy and civil rights group, on Thursday criticized the New York City Department of Transportation's request for information on an artificial intelligence video analysis tool, arguing that the investigation will lead to the expansion of surveillance technology across the city.
The Metropolitan Transportation Authority's request for information, announced Dec. 5, targets AI and computer vision tools that can perform video analytics, such as detecting prohibited objects such as weapons and hazardous materials, recognizing and monitoring abandoned items such as baggage and packages, and predicting, detecting and analyzing abnormal or dangerous behavior, such as spikes in traffic or stampede hazards.
The critical response to the RFI comes after Michael Kemper, the MTA's chief security officer, told the MTA board last April that the agency was “considering” the use of cameras to detect “questionable behavior on subway platforms.”
The RFI is seeking input from “qualified vendors, technology providers, and system integrators to develop proof of value for scalable technology solutions” that can perform real-time analysis of video feeds from cameras installed on MTA public transit vehicles (e.g., subways and buses). The MTA said this information will help assess the “feasibility, scalability and cost-effectiveness” of the technology and may be used to shape procurements.
The MTA said an “unspecified number of technology providers and system integrators” responded to the RFI by the Dec. 30 deadline.
STOP argued on Thursday that more than 15,000 cameras deployed in 472 subway stations will be used to intrusively monitor the public, citing past failures in the city's use of AI video analysis to enhance public safety.
Former New York City Mayor Eric Adams announced last March that the city's police department would test Evolve's AI-powered gun detection scanner on the city's subways. In October, after the pilot program ended, the Legal Aid Society said the program was “objectively a failure”, citing a high rate of false positives. The following month, the Federal Trade Commission announced in a settlement of its lawsuit against Evolv that it had been found to have “deceptively” promoted weapons scanners to schools and other venues.
“The MTA has no right to treat New Yorkers like guinea pigs in an endless AI experiment,” STOP communications director Will Owens said in a news release. “Any AI software that claims to alert you to so-called anomalous behavior is pure pseudoscience and unfairly targeted.” [Black, indigenous and people of color] and disabled transport passengers. New Yorkers don't have to worry about being cited by law enforcement just because of the way they talk or walk. We already know that the deployment of Eric Adams' AI weapons detector was a fiasco, so the MTA should stop pursuing these surveillance outrages any further. New Yorkers need safe, reliable transportation, not some magical Silicon Valley mentality. ”
In April, the group accused the MTA of admitting last year that it was testing behavioral monitoring AI that scores riders for reasonable behavior. In 2023, the group discovered through a public records request that the MTA had contracted with AI company Awaait to monitor tax evasion on subway fares. The technology used the MTA's vast camera network and Awaait's AI software.
While the MTA claims the RFI is “for informational purposes only and does not constitute a solicitation or commitment to issue a formal implementation agreement,” STOP claims the request will expand the city's oversight capabilities.
The RFI goes on to say that transportation authorities are seeking “solutions that are not only technically robust, but also designed in collaboration with behavioral experts to ensure the system reflects a nuanced, ethical and human-centered approach to safety.”
