UK government flows sanctions data into AI-powered fraud detection tool

By on 26/03/2024 | Updated on 26/03/2024
Illustration of a man sitting at a computer with a fraud alert on the screen.
Image: Mohamed Hassan via Pixabay

UK and US sanctions data has been added to a recently launched AI-powered fraud detection tool being used by UK authorities to root out criminals.

The government announced earlier this month that its Single Network Analytics Platform (SNAP) – which was launched just months ago by the Public Sector Fraud Authority (PSFA) in partnership with fintech company Quantexa – has been “upgraded with thousands of new sanctions and debarment records to help it detect suspicious networks, activity and users that warrant further investigation for organised crime and sanctions evasion”.

Three new datasets have been added to SNAP in the latest upgrade: 18,000 UK and US sanctioned entities, including those introduced following Russia’s invasion of Ukraine; 1,000 World Bank debarments, that are deemed ineligible for World Bank contracts; and 647,000 UK dormant companies that do not have any income.

“With this new data, public sector organisations will be even better supported to detect fraudulent claims on public funds through contracts, grants and loans,” the government stated in its announcement, adding that the tool will “continue to be regularly updated with new datasets, each increasing the government’s ability to detect fraud committed against the public sector”.

The government also says that it will run a number of AI “discovery projects” during 2024-25 to “identify new ways to detect fraud”. The moves came on the heels of the Spring Budget 2024, which contained a commitment that £34 million (about $43 million) of “new investment” would be spent “to expand the PSFA by deploying AI to help combat fraud across the public sector”.

Register now: Global Government Fintech Lab 2024 – this one-day event in Dublin, Ireland, on Thursday 25 April contains a session on ‘Fighting financial fraud with technology’ | Free to attend for public servants CLICK HERE TO REGISTER

‘Upgrading’ anti-fraud efforts

The PSFA was established in August 2022 as part of the Cabinet Office and HM Treasury. Its current annual budget (from April 2023 to March 2024) is £12.7 million (about $16.4 million) to help UK authorities fight levels of fraud and error in public spending currently estimated to be between £33 billion (about $42 billion) and £59 billion (about $75 billion) every year.

The PSFA had awarded a contract worth more than £3.4 million (more than $4.3 million) to Quantexa almost 13 months ago to use “new data and cutting-edge technology, including AI, to find and prevent more fraud across the public sector”. Its partnership with the London-headquartered company has now grown to £4 million (about $5 million), according to the recent announcement. The contract award notice 14 months ago described Quantexa’s remit as being the “provision of a Single Network Analytics Platform for cross-government/and UK banking sector use”.

Companies House became SNAP’s ‘first user’ late last year, according to a government announcement on the topic of anti-fraud in November 2023. This announcement stated that SNAP itself had launched the same month (November 2023).

In its own announcement earlier this month, Quantexa states that its AI tech, “which analyses millions of pieces of public sector data”, will be used “as part of the government’s upgrade[d] efforts to detect improper claims on public funds through mechanisms like contracts, grants and loans”.

In addition to mentioning its use by Companies House, it adds that the technology – which it describes as the “SNAP 2 Platform” – will also enable the Financial Conduct Authority (the UK’s financial services watchdog) “to identify shell companies and bad actors”.

Read more: Fraud fighters: public authorities look to fuse technology with data to battle criminals 

AI to tackle ‘phoenixing’

In addition, the government is also planning to run a number of what are described as “AI discovery projects” in 2024-25 to “identify new ways to detect fraud”, according to this month’s SNAP-focused announcement.

The announcement states that the first of these projects will seek to use AI to identify entities registering and bankrupting successive companies to avoid paying debts – a practice known as ‘phoenixing’.

All the projects will be “rigorously tested” by the PSFA in line with the government’s Generative AI Framework, according to the announcement. The framework, which was published on 18 January, is built around ten core principles for generative AI use in government and public sector organisations.

As well as including the commitment of £34 million to the PSFA, the Spring Budget also contained a commitment to “more than double” the size of the government’s Incubator for Artificial Intelligence, known as ‘i.AI’.

Billed as an “elite team of highly empowered technical experts at the heart of government”, i.AI has been set up to deliver “GenAI-as-a-service, for the whole of government”. The website of i.AI states that government AI efforts are “currently impeded by duplication, lack of infrastructure and issues with scalability, causing delivery blockers or outright failure”.

Stay up to date: Sign up for Global Government Forum’s monthly AI Monitor newsletter

‘Harnessing data and tech more effectively’

A Government Counter Fraud Function Strategy 2024-27 was also published earlier this month (14 March). The ‘Counter Fraud Function’, established in 2018, spans about 13,000 public servants across government (of whom about 11,200 work for the Department for Work & Pensions or HM Revenue & Customs).

The document sets five main objectives, including “harnessing data and technology more effectively”.

Planned actions include: increasing the number of departments onboarded to SNAP; “increasing the delivered benefits derived from data matching and analytical activity”; developing a common framework for counter-fraud data sharing; and “identifying new opportunities to work in partnership with the private sector”.

2023-2024 PSFA ‘delivery plan’ published in July last year also contained a priority workstream to make better use of technology to “find, stop and recover” fraud. The workstream contained four commitments, with the first being that the PSFA would create and deploy SNAP, onboarding three public bodies and creating five “test models”.

Quantexa announced at the beginning of March that it had appointed Boon Hui Khoo, former president of Interpol and senior deputy secretary at the Singapore Ministry of Home Affairs, to its advisory board. The company completed a $129 million (about £104 million) Series E funding round led by the Government of Singapore Investment Corporation (GIC) last year.

Join Global Government Forum’s LinkedIn group to keep up to date with all the insight public and civil servants need to know.

About Ian Hall

Ian is editor of Global Government Fintech a sister publication to Global Government Forum. Ian also writes for media including City AM and #DisruptionBanking. He is former UK director for the pan-European media network Euractiv (2011-2018), editor of Public Affairs News (2007-2011) and news editor of PR Week (2000-2007). He was shortlisted for ‘Editor of the Year’ at the British Society of Magazine Editors (BSME) Awards in 2010. He began his career in Bulgaria at English-language weekly the Sofia Echo. Ian has an MA in Urban and Regional Change in Europe and a BA in Economics, both from Durham University.

Leave a Reply

Your email address will not be published. Required fields are marked *