Employment claims surge 28% as UK workers use AI to turn on their bosses
Key Points
- Single employment tribunal claims rose 28% between April and June 2026
- Interim relief applications jumped from around 20 a year to 20 a month
- Judges link the rise in complex claims to workers using AI
- The Employment Appeal Tribunal has set new rules for using AI in cases
- Unchecked AI documents risk strike-out and costs orders
Single employment tribunal claims rose 28% in a quarter as workers turned to AI. Ministry of Justice figures covering April to June 2026 showed the jump on the previous quarter. The Employment Appeal Tribunal has now set out rules for anyone using tools like ChatGPT to prepare a case.
Employment judges have also reported more complex whistleblowing and discrimination claims, which they link to workers using AI to represent themselves.
Workers can bring these claims from their first day in a job, because they carry no qualifying service requirement, explain legal experts at Browne Jacobson.
Applications for interim relief have climbed even faster. Tribunals across Great Britain typically received around 20 of these applications a year, but now receive around 20 a month.
Interim relief is an emergency remedy available in a narrow group of unfair dismissal claims, such as dismissals linked to whistleblowing. A successful application can keep a worker’s contract running until the full hearing.
The President of the Employment Tribunal issued guidance on interim relief applications in response, noting that many of them showed signs of AI use.
300-page ChatGPT argument
The Employment Appeal Tribunal set out its rules in Hancox v Sutherland and others, a case where a worker submitted a 300-page skeleton argument written by ChatGPT.
Hancox brought claims including whistleblowing detriment and disability discrimination against four individuals. A tribunal struck out his claim after he ran a LinkedIn campaign criticising them and publicising disclosed documents.
He appealed, but chose not to rely on the ChatGPT document at the hearing. He accepted that he had not checked it for accuracy and that it lacked credibility.
The document also failed to comply with the appeal tribunal’s practice direction.
Deputy Judge Pilgerstorfer ruled that anyone using AI must make sure every document they submit:
- Complies with the tribunal’s procedural rules
- Has passed the most thorough accuracy check they can reasonably manage, so facts, evidence and case references are correct
- Sticks to relevant points and focuses on the strongest arguments without needless repetition
The judge said workers representing themselves carry the same personal responsibility for accuracy as anyone else who submits documents to a court or tribunal.
The ruling also made clear that litigants cannot submit unchecked AI output and leave the other side to check it.
Litigants who ignore the rules risk tribunals giving their documents little or no weight. They also face sanctions including strike-out, unless orders and costs orders.
Existing judicial guidance on AI already tells judges to ask unrepresented litigants whether they have checked AI-generated content for accuracy.
Hancox’s single surviving ground of appeal will proceed to a full hearing.
