Colin Gray, principal consultant at SAS UK, looks at the role of tackling fraud when it comes to helping local authorities get their finances in order.

One in five local councils in England are currently at risk of bankruptcy due to reductions in spending power and the rising costs of essential services, meaning that as many as 63 councils could declare bankruptcy in the next year.

Estimates also tell us that the levels of fraud and error in public spending are between £33bn and £59bn – amounts which are likely to be suffered across central and local government.

The current pressures on public expenditure mean that we can ill afford to continue to lose money to fraud, with housing tenancy and procurement frauds being the biggest drains on the public purse.

Councils are under pressure to fight fraud in the most cost-effective way possible. Typically, they have been slower to digitise their services, but this is an essential first step to ensure they become less vulnerable to fraud, cyber-attacks and scams.

The role of AI in fighting public sector fraud

There is a clear desire from government and organisations to utilise the benefits of Artificial Intelligence (AI) in the detection and prevention of public sector fraud.

Members of the International Public Sector Fraud Forum have recognised that AI presents a huge opportunity for the public sector to tackle fraud. It can be used to safeguard taxpayer money and government resources, by reducing losses associated with a range of fraudulent activities such as procurement fraud, tax evasion, and social welfare fraud.

In March, the government announced its Single Network Analytics Platform had been upgraded with thousands of new sanctions and debarment records to help it detect suspicious networks. In the announcement, it stated that this data would enable public sector organisations to be better supported to detect fraudulent claims on public funds. It will also be running AI discovery projects to identify new ways to detect fraud.

And in the Spring Budget 2024, £34m of new investment was committed to expand the Public Sector Fraud Authority and deploy AI to help combat fraud across the public sector. The speed and accuracy of cloud-based AI can help the government to stay one step ahead of fraud, by identifying deviations from normal patterns before they develop into fraudulent activities where significant funds might otherwise be lost.

Accurate analysis, maintaining transparency and future planning

AI can be used to safeguard taxpayer money and government resources, by reducing losses associated with a range of fraudulent activities.

Analysing vast quantities of data in real-time can reduce workload and enable faster responses to emerging threats, allowing AI to establish baseline behaviours and identify complex anomalies which are often impossible to spot manually.

However, these tools are programmed by humans and require human oversight to function properly, so there is still potential for incorrect or harmful outcomes. If errors do occur – for example, if the data used to train AI contains bias – public trust and confidence in AI and public sector organisations will be undermined.

To ensure that this is maintained, transparency as to how AI technologies operate, and how personal data is used to make decisions, must be available to the public under GDPR – albeit at a high level and not in such a way it would ever provide useful information for fraudsters looking to exploit AI.

The public sector must continue to keep up with increasingly sophisticated methods used by bad actors, who seek to exploit weaknesses or vulnerabilities that might exist. It is highly likely they are also using modern technology themselves, so the government and councils need to have as sophisticated tools at their disposal as fraudsters.

Benefits for local councils

Certain instances have shown that local authorities are capable of tackling fraud – Birmingham City Council saved £25m over a five year period as a result of regular data matching. But local authorities are still encountering barriers to tackling fraud, including lack of incentives, and lack of data and information sharing.

Research of government departments that we conducted in 2022 established key barriers to data sharing. These included shortcomings of technology infrastructure (66%), budget restrictions (45%), integration (38%), and the flexibility of new and existing systems (36%). Local authorities require support from senior stakeholders and those in charge of governance, to overcome these challenges.

According to MPs, the 'out of control' crisis that we are witnessing in local government funding can only be ended by the government providing billions of pounds more to councils. More funding might be part of the answer but if effective fraud-detecting AI systems are rolled out and there is buy-in from all levels of management to tackle the issue, then the level of fraud in the public sector can be reduced. This could save vast amounts of money – protecting future funding, too!

It will enable cash-strapped councils to reap the benefits in the long term – ultimately reducing the number of councils that are at risk of bankruptcy, and helping them become less vulnerable to fraud, cyber-attacks and scams.