New York's MTA Explores AI-Powered Predictive Crime Prevention on Subway Platforms

New York City wants subway cameras to predict 'trouble' before it happens

Dominic Preston is a news editor with over a decade's experience in journalism. He previously worked at Android Police and Tech Advisor. New York's Metropolitan Transportation Authority says it's exploring the use of AI systems for "predictive prevention" of crime and dangerous behavior on the city's subway platforms. MTA chief security officer Michael Kemper said that the agency is "studying and piloting technology like AI to sense potential trouble or problematic behavior on our subway platforms." "If someone is acting out, irrational... it could potentially trigger an alert that would trigger a response from either security and/or the police department," he explained during an MTA safety committee meeting Monday, emphasizing that the police response could come "before waiting for something to happen." "AI is the future," he added, noting that the MTA is "working with tech companies literally right now" to investigate "what would work in the subway system." Kemper didn't detail which companies the MTA is working with, how AI will be implemented, or exactly what sort of behaviour the AI-enabled cameras will be expected to detect. However, MTA spokesperson Aaron Donovan confirmed to The Gothamist that the new system won't rely on facial recognition. "The technology being explored by the MTA is designed to identify behaviors, not people," Donovan said. This isn't the first time the MTA has implemented AI. In 2023 it disclosed that it was using AI-powered surveillance software to track fare evaders on the subway, monitoring when, where, and how most fare evasion takes place.

NYC Bets on AI Surveillance to Clean Up Subways, Predict Criminal Behavior

(Credit: U.Ozel.Images / iStock / Getty Images Plus via Getty Images) All public transit systems are digging themselves out of the tough pandemic years, but New York's Metropolitan Transit Authority (MTA) has had an especially brutal stretch after a string of concerning crimes, and it's now looking to AI to make the subways safer. The MTA is working with AI companies to deploy technology that would analyze security footage as it's being filmed, according to Gothamist. Without using facial recognition, it would identify "problematic behaviors" and scan for "potential trouble." The goal is to predict which riders could be criminals and catch them before they act. "AI is the future," says MTA Chief Security Officer Michael Kemper. "We're working with tech companies literally right now and seeing what's out there right now on the market, what's feasible, what would work in the subway system." The system would issue automated alerts to the New York Police Department (NYPD) if it finds a potential issue. About 40% of cameras on subway platforms are monitored in real time, so this system could do the same for all cameras without needing to staff a human. That will come with the risk of the AI tool falsely accusing innocent riders. "Using artificial intelligence -- a technology notoriously unreliable and biased -- to monitor our subways and send in police risks exacerbating these disparities and creating new problems," New York Civil Liberties Union Senior Policy Counsel Justin Harrison tells Gothamist. "Real public safety comes from investing in our communities, not from omnipresent surveillance." Chinese subways have been using facial recognition for years, such as to take rider payments, according to the South China Morning Post. Russia is currently expanding its facial recognition payment system, Face Pay, to subways in more cities as well, Radio Free Europe/Radio Liberty (RFE/RL) reports. "This system will become yet another powerful instrument in today's Russia for the surveillance and control of citizens," says Russia-based human rights organizer Andrei Fedorkov, according to RFE/RL. New York Governor Kathy Hochul has increased video surveillance on subways since she took office, including installing two subway cameras in every car in 2022, Gothamist reports. The Trump administration is putting pressure on the city to reduce crime drastically, and threatening to pull funding if it does not, the AP reports. The same goes for transit systems in Chicago and Washington, DC. "If people can't go to the subway and not be afraid of being stabbed or thrown in front of tracks or burnt...we're going to pull your money," says US Transportation Secretary Sean Duffy in a March interview on Fox. The MTA has clapped back with some data, claiming crime is down 40% compared to 2020. So far, 2025 has seen "fewer daily major crimes in transit than any non-pandemic year ever," according to John McCarthy, the MTA's chief of policy and external relations.

New York wants to use AI and cameras to detect subway crime before it happens

A hot potato: Another authority is turning to AI in the hope of being able to predict crimes or problematic behavior before they occur. New York's Metropolitan Transportation Authority (MTA) plans to use the technology on the city's subways by analyzing live security footage for potential troublemakers. MTA Chief Security Officer Michael Kemper said (via The Gothamist) that the agency is working with AI companies to develop the software that can analyze real-time subway platform feeds. The plan is that if the software identifies someone who appears to be acting unusually, suspiciously, or irrationally, it would trigger an alert that in turn would trigger a response from security or the police department. Kemper said that the police response would come "before waiting for something to happen," i.e., stopping a crime before it happens, Minority Report-style. "AI is the future," Kemper said during a committee meeting. "We're working with tech companies literally right now and seeing what's out there right now on the market, what's feasible, what would work in the subway system." MTA spokesperson Aaron Donovan emphasized that the predictive system would not use facial recognition, adding that the AI technology is designed to identify behaviors, not people. There have been several instances of unprovoked attacks in New York's subway systems in recent times, including people being pushed onto the tracks. Ten people were murdered in the NYC subways last year, according to police. Not everyone is happy about the planned use of this technology. "Using artificial intelligence - a technology notoriously unreliable and biased - to monitor our subways and send in police risks exacerbating these disparities and creating new problems," wrote New York Civil Liberties Union Senior Policy Counsel Justin Harrison in a statement. "Living in a sweeping surveillance state shouldn't be the price we pay to be safe. Real public safety comes from investing in our communities, not from omnipresent surveillance." This is far from the first crime-predicting technology we've seen. Earlier this month, it was reported that the UK government is developing a homicide prediction algorithm to identify potential violent offenders. South Korea is also testing "Dejaview," an artificial-intelligence platform that scans live CCTV feeds to spot patterns linked to imminent crime and alert authorities before it happens. Back in 2022, an academic team announced it had created a model capable of forecasting crime up to seven days ahead with roughly 90 percent accuracy. That same year, reports emerged that China was exploring large-scale citizen profiling, aiming to use automated analytics to flag individuals who might become dissidents or lawbreakers before any wrongdoing occurs.