Business
AI-Driven R&D Tax Claims Could Be Rejected Without Human Oversight, Warns Expert
Companies relying heavily on artificial intelligence (AI) to prepare their Research and Development (R&D) tax claims could risk having their claims rejected by HMRC if the process lacks proper human oversight, according to a warning from Blick Rothenberg, a prominent audit, tax, and business advisory firm.
Ele Theochari, a Partner and R&D specialist at Blick Rothenberg, highlighted that the UK government’s recently unveiled AI Opportunities Action Plan presents both “opportunities and risks” for R&D claimants. The growing trend of using AI-based tools to compile and submit R&D claims, as well as additional information forms, has raised concerns. Some providers have even falsely claimed to have special privileges with HMRC, Theochari notes.
Theochari expressed concerns over the quality of AI-generated R&D submissions, warning that many claims are “wordy but lack substance.” She emphasized that such claims are vulnerable to scrutiny by HMRC. In fact, some large R&D companies, particularly those focused on high-volume submissions, have already gone out of business in the past four years due to subpar submissions and the inability to defend follow-up investigations.
Although AI can be an asset in streamlining certain aspects of the R&D claims process, Theochari stressed the importance of having a knowledgeable adviser involved in the process. Even when AI is fed with accurate data, it can still produce errors and “AI hallucinations,” which compromise the integrity of a claim. HMRC itself has encountered similar issues when attempting to use AI for fact-checking during compliance queries.
Despite these concerns, Theochari pointed out the potential of AI to enhance the R&D claims process. AI can be effectively used to summarize complex technical information, identify baseline technologies, conduct research, and manage large calculations. However, she stressed that expert input is essential to ensure any AI-generated content is factual, relevant, and ready for HMRC’s scrutiny.
With the increasing reliance on AI in the tax and business advisory sectors, Theochari’s warning serves as a crucial reminder to companies that while AI can offer efficiency, human oversight remains essential for ensuring the accuracy and credibility of R&D tax claims.