Background
The Employment Tribunal struck out Terence Hancox’s claims against four individual respondents after finding that he had conducted the proceedings scandalously, unreasonably and vexatiously through a LinkedIn campaign and that a fair trial was no longer possible. Hancox appealed, and the Employment Appeal Tribunal listed an inter partes preliminary hearing to decide whether the appeal disclosed a reasonable basis for proceeding.
No required supplementary bundle had been prepared before that hearing. Hancox, then acting in person, submitted a 300-page skeleton argument created using ChatGPT that did not comply with the EAT Practice Direction. With assistance from ELAAS counsel at the hearing, he pursued only one ground: whether he had received adequate notice that the strike-out application would be determined and a fair opportunity to prepare and participate. He later sought to revive another proposed ground concerning whether the Tribunal had properly considered his rights under articles 6 and 10 of the European Convention on Human Rights.
The EAT also addressed Hancox’s disclosure request, applications to review earlier orders, repeated procedural noncompliance, and extensive correspondence with the Tribunal.
The Court’s Holding
The EAT held that the procedural-fairness ground advanced at the preliminary hearing was reasonably arguable and should proceed to a full appeal hearing. It did not decide that ground on its merits. All other grounds within the notice of appeal were dismissed, and the EAT refused to let Hancox withdraw from the position taken at the preliminary hearing and pursue the additional human-rights ground.
The EAT dismissed Hancox’s disclosure application and certified it totally without merit because the requested correspondence and information concerning an Environment Agency investigation were irrelevant to the surviving ground. It also dismissed his applications to review previous orders and issued directions governing the bundles, skeleton arguments, authorities and future applications for the full hearing.
The EAT stressed that all parties, including litigants in person, must comply with its Rules and Practice Direction. Anyone using generative AI remains personally responsible for checking that submissions are procedurally compliant, accurate, relevant, concise and non-misleading. The EAT warned that noncompliant skeleton arguments may be rejected or attract other sanctions and that continued excessive email correspondence could lead to restrictions on how Hancox communicates with the Tribunal.
Key Takeaways
- Only the arguable notice-and-participation ground was permitted to advance to a full hearing; the EAT did not overturn the underlying strike-out decision.
- Litigants cannot submit unchecked generative-AI output or shift responsibility for verifying it to an opponent.
- EAT skeleton arguments must comply rigorously with the Practice Direction, including its requirements of concision, accuracy, relevance and clear organization.
- Disclosure irrelevant to the legal issue on appeal may be refused as totally without merit, and repetitive or noncompliant filings may prompt strict case-management controls.
Why It Matters
The decision gives direct guidance on the use of generative AI in Employment Appeal Tribunal proceedings. Although AI may assist represented and unrepresented parties, it does not relax procedural obligations or the duty to verify facts, authorities and legal propositions before filing.
The judgment also illustrates the EAT’s gatekeeping role at a preliminary hearing: an appeal may proceed on a narrowly defined, reasonably arguable error of law while additional grounds, new evidence and collateral applications are excluded. It further signals that the Tribunal may impose firm controls when excessive submissions consume a disproportionate share of judicial resources.