Tackling public sector fraud with AI tools

August 21, 2024

Engineer use combination of AI technology and robotic arms to help with production lines and other activities to ensure accuracy and reduce errors and waste from production — Image: © Aree Sarak | iStock

Keith Williamson and Narisa Gore from Alvarez & Marsal, discuss the increasing challenges of public sector fraud and the potential of artificial intelligence tools to mitigate them

In recent years, fraudsters have undoubtedly become more sophisticated in their use of technology. Scammers are creating more credible schemes to convince victims to part with their money, generating more realistic data trails and documentation to cover their tracks, and facilitating the movement and laundering of fraud proceeds.

No targets have been spared, from individuals to large corporations and public bodies. Recent examples illustrate how fraudsters have deceived victims by convincingly imitating family members, work colleagues, customers, suppliers, law enforcement agencies, government employees, and tax officials via emails, social media messages, phone calls, and video calls.

The impact of fraud on the public sector

Fraud has had a stark financial impact on the public sector in the U.K., particularly since the COVID-19 pandemic. A 2023 National Audit Office (NAO) report revealed that fraud in government expenditure reported in accounts audited by the office amounted to £21 billion in the two years following the pandemic, nearly four times the total of £5.5 billion in the two years prior. (1) Of the £21 billion, it is estimated that £7.3 billion is related to temporary COVID-19 schemes. The UK’s Public Sector Fraud Authority (PSFA) estimates that fraud and error in public spending is currently between £33 billion and £59 billion per year (excluding COVID-19 schemes). (2)

Government steps to counter fraud

However, the government has been taking concrete steps to counter these challenges – both by seeking to recover losses incurred from fraud and error and by preventing and detecting them in the first place. The PSFA, formed in 2022, is the national centre of expertise on the management of public sector fraud and now leads the Government Counter Fraud Function (GCFF).

The GCFP also develops public ’servants’ capability to deal with scamsters. Moreover, the GCFP has introduced a framework for collaboration among counter-fraud experts across the government.

One of the key objectives of the GCFF – and, in turn, the PSFA – is to harness data and technology more effectively. This is to be achieved by measures including but not limited to:

Ensuring key data sets are accessible and promoting data sharing across varying government departments whilst minimising existing barriers to doing so at present and
It is continuing to develop public and private data partnerships to best leverage technology to strengthen fraud prevention and detection products.
This will utilise various technologies, complex data analytics components, and artificial intelligence (AI) tools, including those that deploy machine learning.

AI-powered fraud detection

In 2023, the PSFA launched SNAP (Single Network Analytics Platform), an AI-powered fraud detection tool in partnership with the UK-based company Quantexa. (4) The tool helps detect improper claims on public funds and is regularly updated with new datasets to improve fraud detection in the public sector. It processes billions of data points at high speed to identify suspicious activity. (5)

In addition to SNAP, the government will run a number of AI discovery projects in 2024-25 to identify new ways to detect fraud. The first of these projects will use AI to identify entities registering and bankrupting successive companies to avoid paying debts, which is known as ‘phoenixing.’ (6)

The PSFA has also invested in building additional counter-fraud tools with the government’s Incubator for AI. The AI is a team of technical experts who aim to help departments harness the potential of AI to improve lives and deliver public services. The authority expects that AI will increasingly prove an aid for counter fraud including pre-investigation to:

1. Identify lines of enquiry and during investigation to gather and organise evidence for disclosure;
2. Summarise evidence to inform fraud measurement and risk assessment;
3. Use patterns to detect and prevent fraud and irregularity upstream; and
4. Aid post event assurance. (8)

Looking ahead at public sector fraud

In its 2022-23 annual report, the PSFA reported that it used cutting-edge analytics and technology to find and stop fraud, including £140 million of potentially fraudulent payments in the Bounce Back Loans Scheme. (9) It also announced that it had exceeded its target of £180 million in audited counter-fraud benefits for the period and delivered a total of £311 million in such benefits.

The government, through the PSFA, appears to be taking meaningful steps to protect and recover public funds from the scourge of fraud. Technology is already playing an essential role in this endeavour. In the future, with the right data sets and business rules, AI could even predict fraud risks, which will assist the public sector further in the fight against fraud.