The UK’s tax authority is turning to AI to help identify fraud

HM Revenue & Customs (HMRC) has entered into a 10-year, £175 million agreement with British technology firm Quantexa to deploy artificial intelligence for detecting tax fraud and errors, marking a major shift in the UK government’s approach to tax enforcement.

The partnership, announced in March 2024, will see Quantexa’s AI systems integrate with HMRC’s existing data to identify irregularities in tax filings and payments. The technology aims to uncover both intentional fraud and unintentional errors, improving the accuracy of tax collection and compliance.

Quantexa, founded in London in 2016, will combine its data analytics tools with HMRC’s collected data and other sources to support investigations and customer service efforts. The company emphasized that its AI tools are designed to support human decision-making, not replace it, with decisions subject to human review.

Why It Matters

This development signifies a major adoption of AI technology in government tax agencies, aligning the UK with global trends. It could improve the efficiency of fraud detection, reduce revenue losses, and enhance taxpayer compliance. However, it also raises questions about data privacy, transparency, and the accuracy of AI-driven investigations.

Background

In recent years, governments worldwide have increasingly incorporated AI into financial and legal enforcement. The US Treasury Department announced in 2024 that it used AI to prevent over $4 billion in fraud and recover payments. The UK’s move with HMRC reflects a broader trend toward digital transformation in public sector fraud detection efforts.

“In government environments, AI cannot operate as a black box. Decisions need to be transparent, auditable, and explainable, particularly in areas affecting citizens directly.”

— Vishal Marria, CEO of Quantexa

“We will never take HMRC data away from the HMRC environment to ensure data security.”

— Vishal Marria, CEO of Quantexa

What Remains Unclear

It is still unclear how effective the AI tools will be in practice, how often false positives or negatives may occur, and how taxpayers will perceive the increased use of AI in tax enforcement. The long-term impact on taxpayer trust and data privacy remains to be seen.

What’s Next

Next steps include the implementation phase, with HMRC and Quantexa likely to monitor the AI system’s performance over the coming months. Further updates on effectiveness, transparency measures, and potential adjustments are expected as the technology is integrated into routine operations.

Key Questions

How will the AI system improve tax fraud detection?

The AI will analyze large datasets to identify patterns and anomalies indicative of fraud or errors, enabling faster and more accurate detection than manual methods.

Will this AI system lead to false accusations against taxpayers?

The AI findings will be reviewed by human investigators to prevent false accusations. Quantexa emphasizes that its technology supports, not replaces, human decision-making.

How secure is the data used in this AI system?

Quantexa has stated that it will never transfer HMRC data outside of the HMRC environment, ensuring data remains secure and under government control.

Could this technology be expanded to other government areas?

While currently focused on tax, similar AI applications could be adopted in other public sector areas such as social benefits, customs, or law enforcement, depending on outcomes and policy decisions.