The HMRC faces scrutiny over its use of AI in assessing R&D tax relief claims. A tribunal has ordered HMRC to disclose whether it uses AI in these assessments, following concerns raised by tax advisors who have noticed anomalies in HMRC's correspondence, such as Americanised spellings and nonsensical content. These advisors suspect that AI may be generating some responses.
The core issue revolves around transparency and potential misjudgements by AI. The HMRC initially refused to confirm or deny AI usage, citing that it could help fraudsters exploit the system. However, this argument was overturned by the tribunal, which prioritised public interest and transparency. The decision has sparked debate about the balance between protecting taxpayer data, preventing fraud, and ensuring fair assessment of R&D claims.
The tax authority must reveal its position, potentially reshaping interactions between taxpayers, advisors, and the UK's innovation incentives. The ruling could have broader implications for how public authorities use AI in tax disputes and other areas.
