AI hallucinations in employment tribunal claims: How we avoid them

Why WorkDisputes is different

AI can save hours when preparing an Employment Tribunal claim—but only if you can trust the information it provides. Some AI chatbots can invent tribunal cases and legal citations. WorkDisputes helps you research and prepare your claim using your own facts, Tribunal procedure and searchable Employment Tribunal decisions.

What is an AI hallucination?

An AI hallucination is when artificial intelligence generates legal information that appears genuine but is inaccurate or entirely fabricated—for example, a non-existent Employment Tribunal case, an invented quotation or an incorrect legal citation.

Large language models predict plausible text; they do not automatically verify legal authorities. Unless AI is connected to reliable legal sources and used within a guided process, it can confidently produce convincing but incorrect legal information.

The problem is not AI. The problem is unguided AI.

If you're bringing your first Employment Tribunal claim, it's understandable to turn to AI for help. You may have been dismissed, discriminated against, bullied, made redundant or left trying to make sense of a stressful workplace dispute.

AI can explain procedures, organise information and save considerable time. The difficulty is knowing whether the legal authorities it produces are genuine.

A fabricated case or quotation can damage credibility, distract attention from your evidence and create unnecessary work for everyone involved. The problem is not AI itself. The problem is relying on unrestricted AI conversations without checking the underlying legal sources.

The risk is highest when someone asks an open-ended chatbot to "find cases for my claim" and then copies the response into a grievance, ET1, schedule of loss, witness statement or skeleton argument without checking it.

WorkDisputes is not built as a blank chat box. It is designed as a safer way to prepare an Employment Tribunal claim by starting with your facts, checking procedural issues, searching processed Employment Tribunal decision data, presenting labelled case summaries, linking to source material where available and helping you draft around evidence rather than invented authorities.

Why fake AI case law matters

Employment Tribunal claims are usually decided on the evidence, chronology, credibility and the correct application of employment law—not on the number of cases cited.

Depending on your claim, relevant legislation may include the Employment Rights Act 1996, Equality Act 2010, Working Time Regulations 1998 and other employment legislation. Case law can help explain how legal principles have been applied, but inaccurate or fabricated authorities can seriously undermine an otherwise genuine claim.

What fake AI law looks like in a claim

• A case that does not exist but is presented as genuine.
• A real case name paired with a quotation or paragraph that is not found in the judgment.
• A citation from the wrong court, wrong year or wrong area of law.
• An invented section number, regulation or legal test inserted into a polished document.
• A case summary that changes the actual reason the claim succeeded or failed.

Recent examples show the risk is real

Concerns about AI-generated legal authorities are no longer theoretical. Recent judgments from the courts and Employment Tribunal show how fabricated citations can waste time, increase costs and undermine credibility.

1. 01

   6 June 2025 — Ayinde v London Borough of Haringey and Al-Haroun v Qatar National Bank

   The Divisional Court dealt with false authorities linked to suspected or actual AI use. In one matter, a schedule contained 45 citations, of which 18 did not exist. The Court referred lawyers to their regulators, highlighted the risk of wasted costs and stated that the contempt threshold had been met.

2. 02

   2 December 2025 — Mr O Harwood-Allen v London United Busways Ltd

   In an Employment Tribunal reconsideration judgment, the judge was unable to locate one of the cited authorities despite checking legal databases and published reports. The judgment observed that one possible explanation was the use of AI or a large language model.

3. 03

   Judgment dated 18 November 2025, published 2026 — A v British Transport Police Authority

   The Tribunal recorded that authorities supplied by the claimant appeared to result from AI hallucinations. The respondent explained that significant time and expense had been spent checking them, and the claimant apologised and confirmed they would no longer rely on AI-generated background research without verification.

Why this can damage your case

• Credibility: if one authority is fabricated, the Tribunal may view the remainder of the document more cautiously.
• Focus: false citations distract attention from the evidence, dates, witnesses and procedure that usually determine Employment Tribunal claims.
• Leverage: employers may use inaccurate authorities to argue that the claim has been poorly researched.
• Costs: where representatives are involved, inaccurate authorities may expose them to wasted costs applications or regulatory scrutiny.
• Delay: everyone spends unnecessary time checking authorities that should never have been included.

Why WorkDisputes is different from generic AI

Instead of asking users to rely on unrestricted AI conversations, WorkDisputes combines artificial intelligence with employment-specific tools, procedural guidance and searchable Tribunal decision data.

The objective is not to replace legal judgment. It is to help you organise your facts, identify relevant issues and reduce the risk of relying on inaccurate legal authorities.

A safer way to prepare an Employment Tribunal claim

Many people preparing a Tribunal claim follow a process like this:

Generic AI

1. 01

   Ask a chatbot

   Begin with an open-ended request.

2. 02

   Receive a list of cases

   The response may look authoritative even when its sources have not been verified.

3. 03

   Copy them into a draft ET1

   Unverified authorities enter the document.

4. 04

   Spend hours checking whether they exist

   Every citation, quotation and legal proposition then needs manual verification.

5. 05

   Correct mistakes before filing

   Remove or amend anything that cannot be verified.

With WorkDisputes

1. 01

   Describe what happened

   Start with your own facts.

2. 02

   Complete a procedural claim assessment

   Identify relevant procedural issues before drafting.

3. 03

   Compare similar Tribunal decisions

   Review patterns in processed Employment Tribunal decision data.

4. 04

   Review source material where available

   Check the published source before relying on an important decision.

5. 05

   Draft your documents around your own evidence

   Keep your chronology and evidence at the centre of the claim.

6. 06

   Prepare your claim with greater confidence

   Move from assessment to research and structured drafting.

A safer way to use AI for employment claims

• Case Assessment converts your answers into a procedural checklist covering investigation, appeals, qualifying service, grievance steps, written reasons, evidence and other issues commonly affecting Employment Tribunal claims.
• Case Comparison Reports compare your facts against similar Tribunal decisions to identify patterns, strengths, risks and missing evidence.
• Case Law Search allows precise searches by case name or broader plain-English searches such as maternity discrimination, redundancy scoring, whistleblowing or grievance retaliation.
• Search results include summaries, decision dates, outcomes and source links where available, allowing you to check the original decision before relying on it in Tribunal documents.
• The ET1 Builder starts with your timeline, selected claim type and relevant legal authorities. It is designed to remove unsupported citation references rather than invent legal sources.

Understanding tribunal decisions

Employment Tribunal decisions can be valuable for understanding how similar factual situations have been analysed. However, unlike decisions of higher courts, they are generally not binding precedents.

For most litigants, tribunal decisions are most useful as examples that help explain how particular facts have been assessed rather than as authorities to quote extensively.

Every summary should be read alongside the original judgment before being relied upon in Tribunal proceedings.

Who this guide is for

• Employees bringing Employment Tribunal claims without legal representation.
• Litigants in person researching Tribunal decisions.
• Employees comparing AI tools for employment disputes.
• HR professionals wanting to understand AI risks in Tribunal litigation.
• Anyone using AI to help prepare employment-related legal documents.

How that helps a litigant in person

• You do not have to begin with a blank page or vague chatbot prompt.
• You can identify whether your claim is weakened by evidence, qualifying service, limitation, procedure or another issue.
• You can understand how similar cases have been decided before deciding which authorities are genuinely worth reading.
• You can build your bundle around the documents that matter: dismissal letters, grievances, appeal outcomes, policies, payslips, witness evidence and medical evidence where relevant.
• You can move from assessment to research to drafting without losing sight of your own facts.

The WorkDisputes route before you file

1. 01

   Run Check My Claim

   Begin with a procedural audit covering qualifying service, investigations, grievances, appeals, deadlines and evidence gaps.

2. 02

   Generate a comparison report

   Compare your circumstances with similar Tribunal decisions to identify patterns, strengths, risks and missing information.

3. 03

   Use Case Law Search

   Search by case name or describe your situation in plain English to locate potentially relevant Tribunal decisions without relying on AI-generated citations.

4. 04

   Verify before relying

   Summaries are designed to help you assess relevance quickly, but any authority relied upon in Tribunal documents should be checked against the published source wherever possible. Good practice: Before filing an ET1, witness statement or skeleton argument, verify every legal authority against the published judgment or another authoritative source. If a case cannot be verified, do not rely on it.

5. 05

   Draft your documents

   Use the ET1 Builder, grievance letter, appeal letter, Subject Access Request generator, settlement offer, chronology, evidence checklist, schedule of loss, deadline tracker and compensation calculator to produce practical claim documents.

What the platform does not promise

WorkDisputes is designed to reduce the risks associated with unrestricted AI, but it is not a substitute for legal advice.

The platform does not:

• Guarantee that any claim will succeed.
• Replace a solicitor or barrister.
• Remove your responsibility to review documents before filing.
• Suggest that automated summaries are a substitute for reading important source material.

Automated summaries are designed to help you understand Tribunal decisions more quickly. Where a decision is important to your case, you should always check the published judgment before relying on it.

Common questions

Prepare your claim with confidence

You do not need to rely on open-ended AI conversations or spend hours checking invented legal authorities. Start with your own facts.

WorkDisputes helps you:

• Identify procedural strengths and weaknesses.
• Compare similar Employment Tribunal decisions.
• Organise your evidence.
• Draft practical Tribunal documents.
• Prepare your case using an evidence-led process.

Whether you're at the start of your claim or preparing documents for a hearing, the platform is designed to help you work from your own facts rather than unsupported AI-generated citations.

Start with a free Case Assessment and build your claim on evidence, procedure and source-linked decisions wherever available.