NYC using AI to combat subway fare evasion

Artificial intelligence (AI) is helping to track fare evasion in New York City subways, the Metropolitan Transit Authority confirmed on Friday.

The MTA’s May 2023 report on fare evasion said “computer technology” tracked evasions at seven city stations by counting the number of unpaid entries. The AI-assisted tech then compares that number to the number of paid entries already recorded in its system.

More than 50 percent of subway escapes involve people simply walking through emergency gates, the report found. That number is followed by 20 percent who jump or climb the turnstile, 16 percent who simply slip through the gaps and 12 percent who duck under the turnstile.

The new technology — which found that about 16 percent of riders evade fares, a higher rate than measured by human counters — is also used to track how and when evasions occur.

“Since this technology operates 24/7, it also provides data on how to escape and ebb and flow during the day,” the report reads. “The biggest spikes are usually from 3 to 4 pm, with smaller spikes during the morning rush hour.”

The MTA will thus develop – for the first time – a greatly increased ability to focus on avoiding spikes by station, by day of the week, and by time of day. This in turn will support data-driven experiments with new implementation strategies,” said the MTA. “With technology that provides reliable ‘before’ and ‘after’ escapes, it is increasingly possible to test new methods to find what really works.

An MTA spokesperson confirmed to The Hill that the new technology is being used and noted that it does not share information with the New York Police Department.

“The MTA uses this tool to quantify the amount of fare evasion without identifying fare evaders,” said spokeswoman Joana Flores.

NYC’s transportation authority could lose $690 million due to fare evasion in 2022, the report said. Software tests begin in 2020, Flores said.

The authority expects to expand the surveillance technology to about 30 stations by the end of the year, according to the report.

The software was developed by the Spanish firm AWAAIT, originally for the Barcelona subway system. The Barcelona system is advertised to help law enforcement catch evasive arrests — but Flores reiterated that those capabilities are not part of the New York City system.

That data will be used to inform future fare avoidance policy making decisions. The city has given at least $35,000 to AWAAIT since 2021, according to government spending records.