A technological arms race has broken out to combat the fare-dodging crisis plaguing Britain’s railways.

Faced with the prospect of £500m in lost ticket sales, rail operators are embracing artificial intelligence (AI) and so-called machine learning to crack down on freeloaders.

Evasion tactics that once went little further than hiding in a train lavatory to avoid the guard have changed radically since ticket barriers – first trialled on the Tube in 1964 – were introduced at major stations and across commuter networks.

Miscreants are often seen forcing their way through the barriers to avoid paying, a trend recently highlighted by Robert Jenrick, the shadow justice secretary, as he issued a call to arms. However, evasion tactics span well beyond brute force.

The digitisation of ticketing in particular, with the bulk of payments made by bank card, travel card or mobile phone, has opened up whole new avenues for fare abuse and forced operators to respond with tougher safeguards.

Transport for London (TfL^» (Transport for London - about)), which puts its annual losses from fare-dodging at £130m, says that 4.7pc of Tube passengers – almost one in every 20 – skipped paying or stumped up the incorrect fare in the last financial year alone. Yet, across its wider network of the Tube, Overground, buses, Docklands Light Railway and Elizabeth Line, the rate of offending fell from 3.8pc to 3.4pc after fines were raised to £100. TfL aims to bring it down to 1.5pc by 2030 through an intelligence-led approach combined with new technologies and more proactive enforcement.

While offending in London is well short of New York’s 13pc evasion rate, Siwan Hayward, TfL’s director of security, says that fare-dodging “is not a victimless crime”. She says: “It robs Londoners of vital investment in a safe and reliable transport network. The overwhelming majority of customers pay the correct fare, and it’s unfair to those who do that a minority avoid paying.”

The main weapon in TfL’s armoury is its Irregular Travel Analysis Platform (ITAP), which uses ticketing and journey data, passenger information and CCTV^▸ (Closed Circuit Tele-Vision) to identify repeat offenders. The tool is integral for a TfL investigations team that seeks to identify “high-impact offenders”, some of whom cost the network up to thousands of pounds a year in lost revenue.

After studying data, the team liaises with 500 uniformed enforcement officers and plain-clothes inspectors, who can demand proof of a ticket or travel card while staging sting operations in stations, sometimes supported by the British Transport Police.

This crackdown led to TfL securing fines of £400,000 last year after prosecuting 360 of the most prolific offenders. Among the offences deployed by fare evaders are techniques known as zonal avoidance and re-tokenisation. The former, colloquially dubbed “doughnutting”, involves travel through zones that have not been paid for.

In re-tokenisation, or “card tumbling,” the fraudster deletes a virtual card from a digital wallet on a mobile device after a journey to prevent authorisation of the transaction overnight. The card is then reloaded, at which point it is allocated a new tokenised number, preventing it from being blacklisted.

A similar con involves the use of a bank card with insufficient funds to pay for a ticket. Tube gates identify the card as genuine and will let the holder in and out, only for the payment to bounce when requested.

One offender last year used a contactless card to avoid paying the correct fare on 202 journeys, resulting in £1,427 in fines. Other scams include the abuse of Freedom Passes, which offer free travel in London for the disabled and those aged 66 and over.

Meanwhile, outside of the biggest cities, at least 2,000 of Britain’s 2,500 or so stations still lack barriers. Therefore, bosses are required to rely on traditional checks to ensure passengers are paying.

Graham Sutherland, chief executive of FirstGroup, which runs express trains on the West Coast and Great Western main lines, said on-train ticket checks remained vital to prevent “revenue leakage”. But there is no doubt that technology has been a game-changer.

LNER^» (London North Eastern Railway - about), which runs trains between London King’s Cross and Edinburgh, has led the way among long-distance operators in deploying AI to uncover unusual ticket-purchase patterns and identify fraudsters, setting up a “machine learning team” to work alongside its risk experts. A pilot project that began in 2023 immediately identified a customer who had failed to pay for £10,000 worth of travel. Using information uncovered by the new technology, the team then built a case, which resulted in the customer repaying the full loss to LNER.

Paul Larder, at LNER, said the application of machine learning had been revolutionary. He said: “Previously, we’ve relied on our revenue protection team to identify customers who deliberately purchase incorrect tickets for travelling on our services. By using AI, we can accurately analyse large amounts of information quickly and identify patterns that our team can investigate further.”

While companies are ramping up efforts to catch offenders, a report from the Office of Rail and Road this month identified “significant inconsistency” in how operators target fare-dodgers, which has led to unnecessary and unfair prosecutions.

Lord Hendy, the rail minister, who put the cost of fare evasion in England outside London at £400m a year, said a planned new ticketing system as part of the nationalisation programme should reduce instances of people mistakenly buying the wrong tickets. However, for the thugs who push through – known as bumpers, jumpers and tailgaters – sturdier barriers are being created.

Developed by Cubic Transportation Systems, the barriers stretch from the floor to shoulder height and use AI-aided scanning software to detect instances of fare-dodging. Cubic says the gates “can accurately detect, record and flag fare evasion as it’s happening, distinguishing between different types of fare evasion, such as pushing through and climbing under the paddles or tailgating”.

Sir Sadiq Khan’s TfL has not yet said if it will order the equipment. However, if they do, they may finally offer some solace to law-abiding passengers confronted by petty criminals cheating the system while regular staff, who are forbidden from staging physical interventions, look on.

How much is the cost of 100 staff per week being assaulted and having time off work as opposed to the amount of fares being lost.

.....Graham Sutherland, chief executive of FirstGroup, which runs express trains on the West Coast and Great Western main lines, said on-train ticket checks remained vital to prevent “revenue leakage”. .....

Split tickets cause all sorts of challenges. They may offer the traveller a cheaper journey, but they cause all sorts of challenges to the operators. I have heard stories of folk purchasing split tickets for their journey; showing the relevant part when asked to show their ticket; and then applying for a refund for the tickets not checked - and openly boasting about it! That is certainly a form of fare evasion.