HMRC acknowledges the use of AI to monitor taxpayers’ social media

HMRC admits for the first time that it uses artificial intelligence (AI) to monitor taxpayers’ social media accounts as part of a criminal investigation into tax fraud.
The Revenue Agency said that if AI tools are used with the department’s traditional checks to analyze online posts, including posts about expensive holidays or large purchases, their posts if they are inconsistent with a person’s announced income. Officials insist that the technology is only deployed in criminal cases and has “appropriate safeguards” and laws.
Amid Westminster’s growing concern about AI’s expansion in tax enforcement, this disclosure stems from disclosure and fears of wider use of it in the future.
Senior Conservative MP warns that reliance on automation tools can lead to errors due to insufficient human supervision.
“If they start taking legal action against individuals based on this, it seems harsh … without human inspections, you’ll find that there will be problems.”
Sir John Hayes, former security minister and chairman of the Conservative MP group, got similarities with the Post Office Horizon Scandal: “The idea that making machines have to be always the right one is what led to the Post Office scandal. I’m a huge AI skeptical.”
The AI monitoring tool runs with Connect, and HMRC’s data analysis system is used for routine tax investigations. Connect billions of data points introduced more than a decade ago, from bank transactions to property records to potential tax evasion.
Prime Minister Rachel Reeves set a goal to recover £7 billion in the UK’s £47 billion “tax gap”, with HMRC officials last month promulgating a strategy that envisions AI embedded in the “day” tax process.
The department is trialing AI-powered “assistant” to help the public complete tax returns and support compliance personnel to review them. If the pattern in the return suggests false information, the system may issue a warning and if fraud is proven, it may be used later as evidence.
The court ordered the HMRC to reveal that concerns about the role of AI in decision-making have increased before September 18, whether AI is used to evaluate claims for research and development of tax credits. The ruling follows a free information requirement from tax expert Tom Elsbury, who believes that the AI may have determined the outcome of certain claims.
Ministers insist that there is always a human “cycle” in decisions that affect individuals, while the HMRC insists that humans retain “the ultimate statement” in law enforcement actions.
The Ministry of Work and Pensions also tried AI tools, with 20,000 civil servants using the technology to draft documents and summarize meetings. HMRC has approached 12 tech companies to propose using AI to help close £46.8 billion of unpaid taxes – most of which are related to maritime accounts, a government source said.
A HMRC spokesperson said: “Using AI for social media monitoring is limited to criminal investigations and is affected by legal oversight. AI supports our processes but cannot replace human decision-making. More use of AI will allow our employees to spend less time on administrators, more time helping taxpayers, and better target fraud and evasion.”



