In a landmark announcement, the UK government has revealed that a new artificial intelligence tool has been instrumental in recovering nearly half a billion pounds lost to fraud in the last year. This figure, £480 million, represents the largest sum ever reclaimed by government anti-fraud teams in a single 12-month period.
The significant savings, uncovered by the BBC, stem from a multi-pronged approach that cross-references data across different government departments, supercharged by a newly developed AI system. A substantial portion of the recovered funds—£186 million—is related to fraudulent activity that exploited schemes during the Covid-19 pandemic. Additional sums were recouped from areas including fraudulent council tax claims and the illegal subletting of social housing.
From Pandemic Losses to Public Service Gains
The recovery of pandemic-related fraud addresses a long-standing government pledge. However, the £186 million reclaimed is a fraction of the total losses previously acknowledged. Before the 2024 general election, the now-Chancellor Rachel Reeves stated that over £7 billion of public money was lost to fraud during the crisis.
A key success of the initiative involved preventing fraud through early detection. The AI tool helped identify and block hundreds of thousands of companies with potentially fraudulent Bounce Back Loans from dissolving. These loans, designed to support businesses with up to £50,000 during the pandemic, were criticized for inadequate diligence. The Cabinet Office cited one case where an individual invented a company and transferred the loan money to Poland.
Ministers have stated that the recovered funds will be redirected to essential public services, including the recruitment of new nurses, teachers, and police officers.
AI Tool Goes Global Amid Scrutiny
The centrepiece of the announcement is the "Fraud Risk Assessment Accelerator," an AI tool developed by researchers within the Cabinet Office. Its function is proactive: it scans new government policies and procedures for potential weaknesses before they are rolled out, aiming to "fraud-proof" them before criminals can exploit vulnerabilities. The tool was developed specifically in response to the widespread fraud experienced during the pandemic.
In a move signalling international confidence in the technology, Cabinet Office minister Josh Simons will announce that the UK will license the tool for use by allied countries. It is expected that the United States, Canada, Australia, and New Zealand will adopt the technology in some form. Minister Simons stated that these "cutting-edge AI and data tools" are essential to protect public funds and ensure they do not "line the pockets of scammers and swindlers."
Balancing Innovation with Civil Liberties
This expansion, however, is likely to raise concerns among civil liberties and campaign groups who have previously criticised the government's use of AI. The move comes amidst ongoing scrutiny of algorithmic bias in public services.
Last year, an AI tool used by the Department for Work and Pensions to detect welfare fraud was found to exhibit bias based on characteristics such as age, disability, and nationality. A Freedom of Information disclosure to the Guardian revealed an internal admission of a "statistically significant outcome disparity" in a fairness analysis of the system.
A recent report from Amnesty International also criticised the government's "unchecked use of tech and AI systems," highlighting the potential for rights infringements. The government's announcement therefore sets the stage for an ongoing debate: balancing the powerful benefits of AI in safeguarding public money against the imperative to ensure its use is fair, transparent, and free from discrimination.