HM Revenue & Customs (HMRC) has confirmed it is harnessing artificial intelligence (AI) to scrutinise social media activity for clues of tax evasion in a growing number of criminal investigations. The move marks a new digital frontier for the UK’s tax authority and has ignited a storm of debate over privacy, fairness, and the risks of putting machine learning at the heart of tax enforcement.
“Increased utilization of AI will allow our personnel to dedicate less time to administrative tasks and focus more on assisting taxpayers, while also enhancing our ability to target fraud and evasion, thereby increasing revenue for public services,” an HMRC spokesperson stated.
The taxman’s eyes are now on the digital lives of suspected fraudsters, trawling posts for evidence of lavish spending or lifestyle choices that seem inconsistent with declared earnings. Social media content about luxury travel or high-ticket purchases could become grounds for deeper investigations. But HMRC is adamant this approach is reserved for criminal cases with “robust safeguards,” not ordinary taxpayers.
Chris Etherington, partner at RSM UK, urged caution: “AI could assist HMRC in optimizing its operations and facilitate the aggregation of all pertinent information regarding an individual. However, the automation process carries risks, including potential cases of mistaken identity, particularly concerning fake or compromised social media accounts. It is essential that human involvement remains in the process to ensure that sound decisions are reached and all relevant data is considered.”
Bob Blackman, senior Conservative MP, voiced doubts about giving AI too much authority: “If they suddenly start taking legal action against individuals based on that, it seems draconian and very challenging—to put it mildly. You’ve got to have a check and balance. The risk is that AI gets it wrong and someone is pilloried—it seems a bit strange if they start doing that with AI. Without a human check, you can see there’s going to be a problem.”
HMRC stresses employment is not threatened by automation; it plans to hire 5,500 more compliance staff. The agency insists AI will supplement, not supplant, human judgement.
Yet with the government now amending HMRC’s privacy policy – from a guarantee of “human judgement” to mere “human involvement” – critics say the changes could be “the thin end of the wedge.” In an era defined by data and digital footprints, the debate on whether AI should be a state-sanctioned digital informant is only set to intensify.
“Any assistance in this area is likely to be positively received by taxpayers, as HMRC’s current online guidance can often be overwhelming for many,” Etherington added, highlighting efforts to use AI to help taxpayers navigate more than 100,000 pages of guidance on the HMRC website.
As HMRC embraces AI to close its projected £7bn share of the “tax gap,” the government faces the challenge of balancing innovation with vigilant oversight—and answering the public’s hard questions about privacy, fairness, and the power of AI.
