Question to the HM Treasury:

To ask the Chancellor of the Exchequer, what assessment she has made of the extent of the use of Large Language Models in cases of (a) tax evasion and (b) fraud.

HMRC recognises the rapid growth in the use of large language models (LLMs) and other generative AI tools and their potential to affect tax compliance. These technologies are widely accessible and are increasingly used by taxpayers and the wider general public.

Whilst HMRC has seen increasing use of LLMs, both for legitimate and for potentially fraudulent purposes, it has not yet identified evidence of widespread misuse of these tools to commit tax fraud. HMRC continues to monitor this with partners across government and internationally to understand developments and ensure appropriate safeguards are in place.