AI fraud surges as UK public sector lacks defences

UK public sector fraud teams are reporting a rise in AI-enabled scams, while most say they lack the tools and resources required for an effective response.

A survey of 100 UK public sector fraud professionals found that two-thirds had seen an increase in AI-powered fraud schemes. Only 10% said they felt fully equipped with the tools, skills and systems needed to fight fraud effectively.

The research also captured views on the scale of potential savings from stronger controls. UK respondents said, on average, that up to 19% of government budgets could be saved by tackling fraud, waste and abuse.

The study surveyed more than 1,300 public sector fraud professionals across Europe, the Americas and Asia Pacific. The work covered a mix of senior and junior roles responsible for monitoring fraud, waste and abuse.

Capability gap

The survey results indicate a mismatch between perceived opportunity and operational readiness. Nearly every respondent ranked fraud prevention as a top 10 organisational priority. More than a third of UK respondents named it as their number one concern.

Respondents linked the challenge to both financial losses and public confidence in services. A total of 96% of UK respondents said fraud had negatively affected citizen trust. Nearly one in five described the impact as severe.

The National Audit Office has previously estimated the cost of fraud and error to taxpayers at £55-81bn in 2023/24. That equates to around 4-7% of total government spending, according to the figures cited in the study.

Fraud professionals also highlighted the types of activity they encounter most often. Payment fraud and identity fraud were the most common categories experienced by public departments. Respondents expected both to grow over the next five years.

In the UK, the study found that citizen tax fraud ranked as the most prevalent threat after payment fraud. Respondents also flagged procurement fraud and VAT fraud as growing concerns.

AI and analytics

The survey described an environment in which public bodies increasingly face AI-driven fraud. Two-thirds of UK respondents reported an increase in AI-enabled fraud schemes.

At the same time, the survey found meaningful adoption of analytics tools within public-sector fraud teams. In the UK, 40% of departments reported using AI, and 52% said they use predictive analytics.

Respondents reported a range of operational outcomes from these tools. In the UK, respondents said AI already boosts workforce efficiency for 68% of departments. They also said it drives better prioritisation of cases for 47%. They said it improves detection speed by 36% and reduces false positives by 30%.

Planned adoption also appeared high. Nearly all UK departments said they plan to adopt AI or GenAI within two years if they are not already using it. Respondents said synthetic data is the most used today. They also said the anticipated uptake of natural language processing is set to rise fastest.

Barriers cited

Respondents pointed to a set of structural and governance constraints that slow implementation. The top challenge in the UK was responsible use of AI and analytics, at 46%.

Integration with existing systems followed at 43%. Privacy and security concerns came next at 42%. Limited budgets ranked at 35%.

The survey framed those barriers as central to why departments struggle to move from investigation and recovery towards prevention. It also suggests fraud teams see the need for investment in systems and skills alongside technology adoption.

Procurement risk

SAS, which commissioned the research, also pointed to procurement exposure as an area of focus. Based on a large UK government infrastructure project it was involved in, the company estimates that up to 5% of all public sector procurement may be at risk. It put that figure at around £20 billion annually.

The company added that not all of the value at risk is recoverable. It said 1% represents a credible level of realisable savings with the right systems and safeguards in place.

The research sits alongside a broader government push to modernise digital services and explore the use of AI in the public sector. The study referenced the AI Opportunities Action Plan published in January 2025 and a Government Digital and AI Roadmap set for summer 2025.

"Governments are facing a dual challenge: rising fraud threats and falling public trust," said Pete Snelling, Senior Manager Fraud & Compliance, SAS UK. "Our research confirms that AI-driven fraud is growing rapidly, but it also highlights the urgent opportunity to turn the tide. With the right tools and skills, departments can not only prevent losses but restore confidence in public institutions."

The study also found a strong appetite among fraud teams for the wider use of AI, while acknowledging constraints in data, systems, and funding. "There's huge enthusiasm for AI, but also very real obstacles," said Snelling. "Governments want to act, but they're often hamstrung by outdated systems, siloed data and lack of capacity. At SAS, we're focused on closing that gap with powerful analytics that are also accessible and scalable."