Two recent decisions highlight the risks associated with the use of artificial intelligence in litigation and the potential consequences where material generated using AI is advanced without proper verification.

Two recent decisions highlight the risks associated with the use of artificial intelligence in litigation and the potential consequences where material generated using AI is advanced without proper verification.

In Omar Rafique v The Commissioners for HMRC [2026] UKFTT 673 (TC), the First‑tier Tribunal considered an application to reinstate an appeal which had been struck out for failure to comply with directions. In the course of the proceedings, the appellant relied on multiple authorities which the Tribunal found to have been generated by AI and to be either incorrect, irrelevant, or entirely fabricated.

The Tribunal emphasised that there is no prohibition on the use of AI as a drafting tool, but all parties are under a duty to ensure that submissions are accurate and do not mislead the Tribunal. It observed that reliance on fabricated authorities may amount to contempt of court. While no referral was made on the particular facts, the Tribunal made clear that parties who cite non‑existent or mischaracterised authorities risk serious consequences, including potential contempt exposure.

In Anthony Malcolm Cork & Anor v Mark Smith [2026] EWHC 1199 (Ch), the High Court considered misleading material placed before it in correspondence prepared by solicitors at Pinsent Masons LLP. The court found that the material resulted from the use of AI which had not been properly checked.

The court identified the risk that AI tools can generate plausible but incorrect legal analysis, including fabricated statutory provisions and authorities, and reiterated that those conducting litigation remain responsible for verifying all material placed before the court. The conduct in that case led the judge to publicly admonish Pinsent Masons and the relevant fee earners, and the matter was referred to the Solicitors Regulation Authority (SRA). The court noted that a range of responses may be available where false material is advanced, including regulatory action and, in appropriate cases, contempt proceedings.

What this means in practice

These decisions underline that:

• AI is a tool, not a substitute for legal judgment or verification;
• there is a duty on all parties to check the accuracy of authorities and submissions, regardless of how they are generated;
• the risks are not confined to litigants in person — they extend to represented parties and professional advisers;
• advancing inaccurate or fabricated material may give rise to serious procedural and regulatory consequences, including, in some cases, contempt proceedings.

For taxpayers and businesses engaged in disputes with HMRC, the cases reinforce the importance of properly prepared pleadings, evidence and authorities. The use of AI‑assisted material without appropriate review carries material risk, particularly in tribunal and court proceedings where accuracy of citation and analysis is critical.

If your business is involved in a tax dispute, it is important to ensure that all submissions are prepared, evidenced and verified to tribunal standard; not reliant on untested AI‑generated material.

If you would like to discuss an ongoing dispute or require support from specialist practitioners experienced in tax litigation, please contact us.