The Employment Appeal Tribunal gave judgment in Hancox v Sutherland and Others [2026] EAT 139 on 17 September 2026, case number EA-2021-001145-AT, following a preliminary hearing on 29 July 2026. Marcus Pilgerstorfer KC, sitting as a Deputy Judge of the High Court, allowed one ground of appeal to proceed to a full hearing and used the judgment to address the appellant's use of generative AI. The guidance is appellate authority rather than a practice direction or a consultation.
Paragraph 32 of the judgment sets three minimum requirements. Documents must comply with the applicable procedural rules, including paragraph 3.8 of the 2024 EAT Practice Direction on grounds of appeal and paragraph 11.6 on skeleton arguments. Documents must be checked as thoroughly as the litigant is reasonably able, so that the position is not misstated and the tribunal is not misled. Documents must carry only relevant points, focused on the central arguments and put clearly without undue repetition. Paragraph 33 rejects both the submission of raw AI output and any attempt to place the checking burden on an opponent. Paragraph 37 states that litigants must take personal responsibility for any document submitted to the tribunal.
The duties bind represented parties and litigants in person alike, so solicitors, counsel and ELAAS volunteers advising appellants carry them too. The appellant's ChatGPT skeleton was held non-compliant and was not relied on at the hearing. The judgment warns that documents of that kind are likely to be rejected in future and that costs and strike out powers may be considered, which places a verification step inside the drafting routine of any firm filing in the Employment Appeal Tribunal.
The judgment does not impose a duty to disclose that generative AI was used, and sets no page limit beyond the existing terms of the Practice Direction. It draws on R (Ayinde) v London Borough of Haringey [2025] 1 WLR 5147 and on the AI Guidance for Judicial Office Holders of 31 October 2025. A later attempt by the appellant to add a further ground relying on Articles 6 and 10 was refused.
Licentium advises employers, technology providers and professional firms on the use of generative AI in contentious work and on the duties attaching to AI assisted documents. Work we undertake includes AI use policies for litigation teams, verification procedures for tribunal and court filings, and advice on liability where unchecked AI output reaches a decision maker.