Post at your peril: UK’s HMRC uses AI to monitor social media for tax fraud investigations

HMRC has confirmed it uses artificial intelligence to monitor social media posts as part of criminal investigations into suspected tax cheats, marking a significant expansion of automated surveillance in tax enforcement. The technology enables the tax authority to streamline investigations while raising important questions about privacy, accuracy, and the role of AI in government oversight.

What you should know: HMRC, the UK’s tax authority, emphasizes that AI monitoring is limited to criminal investigations and includes human oversight to prevent automated decision-making.

• The technology is only used as part of criminal investigations, not on everyday taxpayers, according to an HMRC spokesperson.
• “Greater use of AI will enable our staff to spend less time on admin and more time helping taxpayers, as well as better target fraud and evasion to bring in more money for public services,” HMRC said in a statement.
• The tax authority has been using AI for several years to support its work, though the social media monitoring capability was first reported by the Daily Telegraph on Tuesday.

How it works: The AI system automates the process of scanning social media posts for evidence of tax evasion, building on years of manual investigation techniques.

• Government investigators have traditionally examined social media posts of suspected tax and benefit fraudsters manually.
• One notable case involved a woman who claimed benefits citing ill health but was exposed through her social media posts about long-distance running.
• HMRC is now partially automating this process through AI while maintaining human oversight of decisions.

Expert concerns: Accountancy professionals warn that AI-powered surveillance carries risks of false identification and requires careful human review.

• “AI could help HMRC to streamline its processes and make it easier to compile all the available information about an individual,” said Chris Etherington, a partner at RSM UK, an accountancy firm.
• “There are of course risks with automating this and there could be cases of mistaken identity, particularly with fake or hacked social media accounts,” Etherington cautioned.
• “There will clearly still need to be a human touch in the process to ensure appropriate decisions are made and all the relevant information is reviewed.”

Broader AI initiatives: HMRC is expanding AI use beyond surveillance to improve taxpayer services and streamline operations.

• The government announced the tax authority will gain 5,500 compliance staff, suggesting the AI implementation isn’t replacing jobs.
• HMRC plans to use AI to help taxpayers navigate over 100,000 pages of guidance on its website.
• The authority is also developing AI tools to help people complete tax returns accurately, as “a lot of tax revenues can be lost through errors and mistakes,” according to Etherington.

Political context: The disclosure comes as the government faces criticism over its broader AI strategy and funding decisions.

• Insiders allege that the government’s national institute for AI is at risk of collapse.
• The technology secretary has threatened to withdraw funding and wants the institute to focus on defense research.
• Despite these challenges, HMRC appears committed to expanding AI use to “make life easier for taxpayers,” Etherington noted.