AI-Generated Grievances and Employment Claims: A New Challenge for Employers?

By Chloe Holmes

The growing use of artificial intelligence (AI) is changing the way individuals engage with employment law. It offers new opportunities to understand workplace rights and draft grievances. However, it also raises questions about the accuracy and reliability of AI-generated legal material.

Whilst AI can be a useful tool to assist individuals who otherwise would not have been able to access legal assistance, relying on AI to produce legal arguments and documents can create significant challenges.

The use of AI in an Employment Tribunal claim has also attracted interest. In particular, the Employment Appeal Tribunal’s (“EAT”) recent comments on a 300-page submission generated using ChatGPT in Hancox v Sutherland and Others [2026] EAT 139 have drawn attention.

Chloe Holmes, associate at Redmans Solicitors, explores the recent use of AI and how employers should respond to AI-assisted grievances and claims fairly.

The 300 Page Submission in Hancox v Sutherland and Others

In the recent case of Hancox v Sutherland and Others, an employee was appealing the decision made by the Employment Tribunal (“ET”) to the EAT. His claim had been struck out by the ET on the basis that he had conducted the proceedings in a way which was “scandalous, unreasonable and vexatious”. He appealed the decision and the case was listed for a preliminary hearing by the EAT.

Before the hearing, the employee submitted a 300-page (132,000 word) “skeleton” argument, created using ChatGPT. The EAT pointed out that the document did not comply with the Practice Direction (the procedural rules and requirements for the court) and that all parties (whether they have representation or not) must comply with the Practice Direction. Since the skeleton arguments did not comply, it was rejected. The EAT judge commented that the document “served to obscure rather than to illuminate”.

The EAT issued a reminder that everyone using generative AI must take personal responsibility to ensure:

1. The documents comply with the applicable procedural rules,
2. The documents have been checked thoroughly (in so far as is possible), ensuring that the content does not mislead the court, and
3. The documents only contain relevant points and are presented in a way which is easily understandable and avoids repetition.

Whilst the EAT recognised that generative AI has the potential to assist those who would otherwise not have access to legal advice, it also recognised the risk of using AI. Such output can be inaccurate, unfocused, excessive and may not comply with procedural requirements.

The EAT also confirmed that it is not acceptable to submit documents using AI without checking them. It is also unacceptable to place the burden on the other party to check them.

Additionally, the EAT recommends that skeleton arguments be between 5-15 pages, with an upper page limit of 20. Anything above that, is likely to be rejected by the EAT in the future.

Can Using AI Make Employment Rights More Accessible?

Despite the concerns highlighted by the EAT, AI has the potential to improve access to employment justice. It can help individuals understand their rights and navigate processes that might otherwise be difficult or intimidating.

Employees raising grievances may be distressed by their experiences at work and may struggle to organise their concerns into a clear, structured account. AI can help employees prepare a chronology of events, summarise relevant correspondence and identify the key issues they wish to raise. This may help employers understand the complaint more readily and facilitate a more focused investigation.

AI may also be particularly beneficial for individuals in the UK whose first language is not English. It may enable them to translate their concerns into English, improve the clarity of their written communications and navigate unfamiliar legal terminology. In this respect, AI has the potential to reduce language barriers that might have otherwise prevented employees from articulating their claims or grievances effectively.

However, accessibility is only beneficial if the information produced is reliable and the resulting document accurately reflects the employee’s circumstances.

The Risk: Hallucinations, Misinformation and Disproportionate Drafting

The major risk of using AI is that it can produce convincing, but ultimately incorrect answers. An AI tool works from the information it is given via the user in the prompt. It does not know what the employee experienced unless that information is accurately provided. An employee relying on the output of AI may find themselves pursuing an incorrect legal argument or misstating the facts.

When drafting legal documents, AI may introduce cases or legislation that do not exist, also called hallucinations. AI hallucinations occur because AI works by identifying patterns and predicting what response is likely to fit those patterns. It may generate plausible-sounding responses, but it does not know the difference between what is correct and incorrect. In tests conducted by OpenAI, it was found that its o3 reasoning model hallucinated in 33% of questions asked during a test about public figures.

Other risks of using AI include the fact that AI outputs are often disproportionately lengthy and repetitive. Irrelevant material can make it harder to identify the core complaint/legal arguments. A grievance or legal claim is not strengthened because it is longer or “sounds more legal”.

Finally, there is a risk that information may not be held securely by third-party AI tools. Employees often enter sensitive information without understanding how that information is stored. Similarly, employers that upload staff grievance records or investigation documents into third-party tools may also face risks. The same applies to legal professionals and anyone bringing Employment Tribunal claims.

For example, a recent judgment from the Upper Tribunal Immigration and Asylum Chamber held that uploading confidential documents into an open-source AI tool such as ChatGPT could waive legal privilege and breach client confidentiality. This is because it effectively puts information online and into the public domain.

How Should Employers Respond To AI-Assisted Grievances and Claims?

For employers, the increasing use of AI in workplace disputes presents a practical challenge: how to address potentially lengthy or inaccurate submissions without losing sight of the underlying grievance and the need for a fair process.

Employers should not dismiss a grievance simply because it appears AI-generated. The employee could still have legitimate concerns, even if the document is difficult to decipher. Where a grievance is lengthy or unclear, employers should try to identify the key issues and ask clarifying questions (where proportionate and appropriate). Employers should also work out which facts need investigating and take steps to verify any evidence provided by the employee.

Any grievance should still be handled in accordance with the Acas Code of Practice on disciplinary and grievance procedures and the Company’s own procedure. The use of AI should not be a reason to delay investigation or deny an employee an opportunity to voice their grievances.

This also applies to Tribunal claims. Employers and their legal representatives should focus on the relevant issues and evidence, comply with procedural requirements, and challenge submissions where they appear to be inaccurate or irrelevant through appropriate channels.

Should Employers Introduce Policies Regulating AI In Workplace Disputes?

In addition to responding appropriately to AI-assisted grievances, employers should consider whether clear workplace policies are needed to regulate the responsible use of AI by employees and HR teams. Employers may wish to consider addressing AI use within their existing acceptable use or grievance policies. Possible policy provisions could cover:

• Permitted use of AI – Explaining that employees may use AI to help organise thoughts, improve clarity or understand workplace procedures, but making it clear that employees remain responsible for the factual accuracy of allegations and documents they submit, including AI-assisted material.
• Confidentiality – Employers may wish to prohibit uploading personal data, confidential employment records or third-party information into unapproved AI systems. This is subject to the employer’s authorised arrangements and applicable data protection requirements.
• Fairness and transparency – Establish appropriate controls for AI use in employment decisions and explain how employees can raise concerns about its use.

Policies should be proportionate, accessible and consistently applied. An outright ban on AI-assisted grievances could disadvantage employees who rely on these tools to communicate effectively.

Ultimately, the challenge for employers is not to prevent the use of AI in employment disputes, but to ensure that it facilitates the fair and effective resolution of workplace grievances rather than complicating the process.

About the Author

Chloe Holmes is an Associate in the Employment Law team at Redmans Solicitors. She supports clients on a wide range of employment matters—both contentious and non-contentious—including assessing claims, drafting legal documents, and advising on workplace disputes.