TLDR: A recent tribunal ruling mandates HM Revenue & Customs (HMRC) to disclose its use of artificial intelligence in decisions regarding Research and Development (R&D) tax credit claims. This follows a Freedom of Information request by tax expert Tom Elsbury, who suspected AI was being used to reject claims. The decision overturns previous refusals by HMRC and the Information Commissioner’s Office, emphasizing public interest in transparency.
In a landmark ruling on August 8, 2025, a UK First-tier Tribunal has ordered HM Revenue & Customs (HMRC), the nation’s tax agency, to reveal the extent of its involvement of artificial intelligence (AI) in decisions concerning Research and Development (R&D) tax credit applications. This significant decision underscores a growing demand for transparency in the use of AI within governmental processes.
The case originated from a Freedom of Information (FOI) request filed in December 2023 by Tom Elsbury, a prominent tax expert and founder of the specialist tax relief software platform, Novel. Elsbury, along with several other tax advisers, grew suspicious of HMRC’s decision-making process, particularly noting the language and structure of rejection letters for R&D tax credit claims, which suggested the deployment of AI tools. His FOI request sought confirmation on whether and when AI was being utilized in these critical determinations.
Initially, HMRC refused Elsbury’s request, citing concerns that disclosing such information could potentially assist individuals attempting to make fraudulent tax relief claims. This refusal was subsequently upheld by the Information Commissioner’s Office (ICO) in 2024. However, Judge Alexandra Marks, presiding over the first-tier tribunal, found Elsbury’s arguments for transparency ‘compelling’ and ruled that the original decisions by both HMRC and the ICO should be overturned. The tribunal explicitly stated that ‘the balance of the public interest lies in disclosing the information requested.’ HMRC has been given until September 18 to comply with this order.
Elsbury has been a vocal advocate for greater transparency, emphasizing the public’s right to know ‘if AI is concluding or forming a decision in tax enquiries’ related to enterprise tax credits, and whether penalties are being issued based on AI-generated decisions. He highlighted the potential for ‘catastrophic consequences’ if sensitive R&D details, particularly those involving defense and military innovations, are processed by opaque AI systems. Such lack of transparency, he argued, ‘undermines taxpayers’ trust and confidence in HMRC’s treatment of claims.’
Also Read:
- States Forge Ahead with AI Governance Amidst Federal Regulatory Vacuum
- Artificial Intelligence Reshapes Corporate Hierarchies, Bridging C-Suite and Workforce
R&D tax credits, introduced by the Blair government in 2000, are designed to stimulate innovation in small and medium enterprises (SMEs) by supporting projects that ‘seek an advance in a field of science or technology by resolving scientific or technological uncertainty.’ The scheme has recently been subject to a crackdown by HMRC following a spate of fraudulent applications and grants made in error. This tribunal ruling marks a pivotal moment in the ongoing debate about the ethical and practical implications of AI integration into public administration, particularly where it impacts financial decisions and public trust.
