The use of generative artificial intelligence (‘GenAI’) in litigation is no longer hypothetical (see our earlier article).
Courts and tribunals are increasingly encountering documents prepared using GenAI tools. They are having to balance the potential benefits GenAI can bring for access to legal assistance against the real practical risks of using GenAI, including inaccurate output, unfocused or excessive answers, or that it might not comply with procedural requirements.
The Employment Appeal Tribunal’s (‘EAT’) recent decision in Hancox v Sutherland and others [2026] EAT 139 provides useful commentary on the responsible use of GenAI in employment tribunal litigation. Importantly, the EAT did not suggest that litigants should avoid using AI. Instead, its judgment emphasised that responsibility for documents produced with the assistance of AI remains with the person putting them forward.
What Happened?
The case concerned an appeal by a litigant in person whose multiple claims against four individual respondents had been struck out by the Employment Tribunal. The Employment Tribunal had concluded that the claimant had conducted proceedings in a scandalous, unreasonable and vexatious manner through a social media campaign on LinkedIn, and that a fair trial was no longer possible.
The claimant appealed and the matter was listed for a preliminary hearing. Ahead of a preliminary hearing in the EAT, the claimant filed a 300-page skeleton argument with a word count of almost 132,000 words. The EAT commented that the skeleton argument “did not come close to complying with the requirements of the PD [EAT Practice Direction]” and, as the EAT observed, the document “served to obscure rather than to illuminate” the case during preparation for the hearing. The claimant explained at the hearing that he had used ChatGPT to create the document so as to prepare it quickly. Ultimately, the claimant did not rely on it.
In addition, the claimant submitted the skeleton argument with a covering email, which acknowledged that the judge could not reasonably be expected to read the entire document before the hearing and invited the respondents to identify factual inaccuracies, incorrect quotations, dates, bundle references or misrepresentations within it. The EAT considered this to be unacceptable.
The EAT also noted that before and after the hearing, the claimant had also made numerous, often repetitive, applications and sent excessive correspondence to the EAT.
The EAT’s Response
The EAT acknowledged that GenAI has the potential to assist people who would not otherwise have access to professional legal advice. At the same time, it identified the risks associated with GenAI use, including that AI-generated output may be legally or factually inaccurate, unfocused, excessive or non-compliant with procedural requirements.
However, the EAT did go on to provide some helpful commentary for parties using Gen AI in litigation:
- Personal responsibility: Drawing in particular on the recent case of R (Ayinde) v London Borough of Haringey [2025] 1 WLR 5147, the EAT said that all litigants and their representatives (including litigants without representation) using GenAI to prepare material for the EAT must take personal responsibility for it. At a minimum, they must ensure that documents:
- Comply with the applicable procedural rules (for example any relevant practice directions);
- Have been checked, as thoroughly as the litigant or representative is reasonably able for accuracy, ensuring that the factual, evidential and legal position is not misstated and the EAT is not misled; and
- Contain only relevant points, with the focus being on the central or best arguments, presented in an easily comprehensible manner and avoiding undue repetition
The EAT was particularly clear that responsibility for checking AI-generated material lies with the individual relying on the document:
“It is not acceptable to submit documents created by generative AI without them being checked, or to attempt to place the onus for checking them on the opposing party.”
Where there are concerns that appropriate checks have not been undertaken, judicial enquiries and potentially sanctions may follow.
- Compliance with the rules of procedure: The EAT explained that compliance with the EAT’s procedural requirements is mandatory, not optional and remain applicable to litigants in person
- Excessive correspondence: Even where AI-generated material is not demonstrably inaccurate, GenAI can make it considerably easier to produce large quantities of apparently sophisticated legal material. The EAT noted that the excessive volume of correspondence and applications in the case had “not advanced the overriding objective” and had required a disproportionate share of the EAT’s resources
It is also worth noting that the claimant succeeded in establishing that one ground of appeal was reasonably arguable and should proceed to a full hearing. The judgment should not therefore be understood as suggesting that the use of GenAI undermines the merits of a litigant’s case. Its focus was on the appropriate use of GenAI and the responsible conduct of litigation.
What Does this Mean for Employers?
Employers are increasingly likely to encounter lengthy, repetitive or unfocused GenAI assisted correspondence, both before and during employment tribunal proceedings. Some tribunal case management orders are also beginning to ask parties (on both sides) to confirm whether GenAI, has been used and whether resulting documents have been checked for accuracy and are focussed on relevant issues identified in the claim.
Given this new GenAI world in litigation, employers may wish to:
- Identify the core issues early: Clarifying the employee’s main concerns at the outset can help keep correspondence and subsequent proceedings focused on the matters genuinely in dispute
- Respond proportionately: Receiving lengthy GenAI created correspondence does not necessarily require an equally lengthy response. Employers may consider responding strategically rather than answering to every allegation or peripheral issue
- Use case management tools where appropriate: If correspondence or applications become excessive due to GenAI’s ability to create vast documents at pace, employers may wish to consider whether existing tribunal procedures, rules and case management directions can help control the volume
- Apply the same principles to employer’s use of GenAI: The EAT in this case puts the onus on the individual relying on the documentation. Any documentation being relied upon should be checked for accuracy, relevance and procedural compliance. Those using GenAI in connection with disputes should also remain mindful of separate confidentiality, privilege and data-protection considerations when deciding what information can appropriately be provided to particular GenAI systems. See our previous article
Strong grievance policies and procedures may also help to identify and narrow down the real issues and concerns being raised by employers and seek to resolve them before they develop into litigation.
