HM Revenue & Customs (HMRC) in the UK has issued guidance for software developers on the use of generative AI (genAI) in software products that support the submission of tax returns and other information to the department.

The guidance is aimed at commercial tax software providers and sets out what HMRC considers to be appropriate use of genAI in tools used by taxpayers.

Access deeper industry intelligence

Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.

Find out more

HMRC says it recognises the opportunities that the technology offers and “encourages innovative use of AI in tax software products”.

However, it stressed that tax software using genAI should be transparent, with developers ensuring that users know when the software is “enhanced by generative AI”.

The user should also be made aware of its capabilities and its limitations.

Additionally, the department expects tools that help prepare tax returns, including those that generate information for users to file an accurate return, to rely on data sources that are “high quality, reliant and reliable”.

GlobalData Strategic Intelligence

US Tariffs are shifting - will you react or anticipate?

Don’t let policy changes catch you off guard. Stay proactive with real-time data and expert analysis.

By GlobalData

HMRC specifies that these data sources include current HMRC guidance, legislation and case law.

The guidance also stresses that systems must remain subject to human control.

The software should prompt users to review AI-generated outcomes, particularly in “areas involving nuanced tax rules, complex scenarios or specific guidance”, and should remind users that they retain responsibility for the accuracy of their returns.

HMRC emphasises that the software “should support, not replace, human judgment”.

The department also identified security as another key requirement. HMRC expects strong data protection and privacy measures, noting that these tools may handle “highly sensitive personal, commercial and financial data”.