New York City is using AI to measure how much people hate paying for subway tickets

A hot potato: AI algorithms can do much more than generate garbled, unreliable text snippets and eerily repulsive synthetic images. In fact, the Metropolitan Transportation Authority (MTA) is already employing the tech to understand how much fare evasion is impacting the New York City subway service.

According to the May 2023 report published by the MTA on ticket evasion, people unwilling to pay their fair share to use New York’s subway system are costing the city nearly $700 million in lost revenues. This fare evasion is essentially threatening the future of the public transit system, the MTA said, which in turn would “tear at the social fabric” of the most densely populated urban conglomerate in the US.

Numbers don’t lie, and they come from an AI system created by the Spanish company AWAAIT. The software was originally developed for the Barcelona subway network, but it is seemingly working well enough for New York as well. MTA began testing AWAAIT’s tech in 2020, and it is now employing it in seven different subway stations to track fare evasion on a continuous basis.

The AI-assisted software compares the number of unpaid entries with the number of paid entries that were previously recorded by the system. The technology can operate 24 hours per day, every day of the week, MTA said, so it can also provide detailed data about the time of the day when the largest spike in fare evaders is recorded.

The largest number of non-paying passengers is seemingly detected between 3 and 4 PM, MTA said, with smaller spikes happening during the morning rush hours. More than 50% of fair evasions consist of people simply going through the emergency gate, the transportation authority stated, while 20% of evaders prefer to jump or “climb” over the turnstile. Furthermore, 16% of passengers are skinny enough to slip through the gaps, and 12% duck under the turnstile.

Despite being a “threatening” phenomenon for the future of New York’s most popular transportation system, the MTA remarked how fare evaders are not individually tracked by their AI-assisted system. The technology is used to quantify the amount of fare evasion and it does not provide any information for NYPD investigations, the authority said.

Nevertheless, MTA is planning to expand the use of AI algorithms to monitor fare evasion in 30 more subway stations by the end of the year. Data provided by the aforementioned algorithms will be employed to inform future policy-making decisions against artful ticket dodgers. The Barcelona subway system is already using AWAAIT’s AI to assist law enforcement authorities in catching fair evasion, but New York is seemingly uninterested in this additional capability.