Get all your news in one place.
100’s of premium titles.
One app.
Start reading
Evening Standard
Evening Standard
World
Ross Lydall

London Underground: Artificial Intelligence software is being used to detect fare dodgers

Artificial intelligence is to be trialled on the London Underground in a bid to stem soaring rates of fare dodging, it can be revealed.

Computer technology that can identify passengers likely to jump ticket barriers has been tested at Willesden Green station, on the Jubilee line.

Transport for London now plans to expand the trial after successfully linking the AI software with its network of station cameras.

Siwan Hayward, TfL’s director of security, policing and enforcement, insisted the technology was not being used as a “facial recognition tool”.

She said: “The pilot project was able to detect fare evasion attempts through the gate-line and enrich our data and insight on fare evasion levels and methods.

“Following a review by our safety, health and environment team, we will be progressing this to an in-station trial to monitor its effectiveness.”

Latest figures estimate that TfL loses more than £130m a year in income due to fare dodging. About 3.9 per cent of journeys are unpaid – about one in 25.

More than 31,000 penalty fares have been issued since April and more than 11,000 cases are being considered for prosecution.

TfL wants to increase the penalty fare for failing to pay the correct fare to be increased from £80 to £100 but says this has yet to be approved by Mayor Sadiq Khan, despite being first proposed a year ago.

TfL declined to comment on the AI trial. But it is understood that the project at Willesden Green used AI algorithms and motion detection to “detect the act of fare evasion” – passengers passing through barriers without paying.

TfL has known for some time that the “wide aisle” gates – which allow access to wheelchair users, people with children in prams and pushchairs and passengers with luggage – are often targeted by fare dodgers.

The gates can be easily pushed open and also close slowly – meaning a fare evader can “tailgate” behind a paying passenger.

TfL will use the information from the AI pilot to design more secure “wide aisle” gates, and to improve its existing “irregular travel analysis platform”, which gathers data on fare dodgers.

Figures presented to TfL’s customer service panel meeting tomorrow show there has been a 26 per cent increase in penalty fares being served on fare dodgers between April and September, compared with the same period last year.

A total of 213 people – thought to be hard-core offenders - have been stopped and interviewed. Each had evaded an average of 89 journeys – worth £821 in unpaid fares.

However TfL has a shortage of almost 200 enforcement officers. Only 452 of the 551 posts are filled – and of those at work, 81 have been seconded onto other duties, leaving only 371 available to catch fare-dodgers.

TfL has also had to contend with controversies over some "heavy handed" enforcement operations, including when a young mother was apprehended in Croydon.

This resulted in a Met police officer being placed under under criminal investigation after the woman was found to have been wrongly arrested.

Fare dodgers are known to present a wider problem – they are responsible for more than half of the verbal and physical attacks on Tube station staff.

Attacks have soared by 50 per cent, with 647 incidents reported in the last six months.

The Jubilee line is the second worst on the Underground, after the Northern line, for crime.

The Standard revealed last week that Tube crime was up 56 per cent year on year, fuelled by an 83 per cent increase in thefts and a 107 per cent increase in robberies.

Sign up to read this article
Read news from 100’s of titles, curated specifically for you.
Already a member? Sign in here
Related Stories
Top stories on inkl right now
One subscription that gives you access to news from hundreds of sites
Already a member? Sign in here
Our Picks
Fourteen days free
Download the app
One app. One membership.
100+ trusted global sources.