The UK government has recovered almost £500m over the past year in what officials describe as the most successful anti-fraud operation to date, the BBC has learned. The gains were achieved through a combination of data cross-checking between departments and the use of a new fraud-detection system that will now be licensed internationally.
The Cabinet Office confirmed that £480m was reclaimed between April 2024 and April 2025, marking the largest single-year recovery by government anti-fraud teams. More than a third of the figure — £186m — was linked to fraudulent claims made during the Covid-19 pandemic, including abuses of the government’s Bounce Back Loan scheme.
Launched during the pandemic to support struggling businesses with loans of up to £50,000, the programme has faced heavy criticism for weak safeguards that allowed opportunists to exploit the system. Officials said the latest recovery effort blocked hundreds of thousands of companies with potentially fraudulent loans from dissolving to avoid repayment. In one case, investigators uncovered a fictitious business created by a woman who later transferred loan money abroad.
The Cabinet Office says the recovered funds will be directed towards frontline public services, including the recruitment of nurses, teachers, and police officers.
At the centre of the effort is a new system called the Fraud Risk Assessment Accelerator, developed by Cabinet Office researchers. The tool analyses new government policies and procedures for potential vulnerabilities before they can be exploited, with officials claiming it could “fraud-proof” policies at the design stage.
Cabinet Office minister Josh Simons is expected to formally announce the savings at an international anti-fraud summit on Wednesday, hosted jointly with the US, Canada, and Australia. He described the tool as part of “cutting-edge AI and data solutions” ensuring public funds are used for essential services rather than “lining the pockets of scammers and swindlers.”
The UK will now license the system abroad, with the US, Canada, Australia, and New Zealand among the first countries expected to adopt it.
However, the expansion of government use of automated fraud detection tools is not without controversy. Civil liberties campaigners have previously criticised similar systems for potential bias and lack of accountability. A Department for Work and Pensions tool used to detect welfare fraud was found last year to produce “statistically significant outcome disparities” related to age, disability, marital status, and nationality, raising concerns about fairness.
Earlier this year, Amnesty International warned of the UK’s “unchecked use of tech and AI systems” in public policy. Campaigners are likely to scrutinise the rollout of the Fraud Risk Assessment Accelerator both domestically and abroad, despite its apparent success in recovering public funds.
