AI-drafted claims push UK employment tribunals to 64,000
The Economist reports UK employment tribunals are buried under chatbot-drafted claims, with open cases up 55 percent to 64,000 and non-existent statutes cited.
The numbers
Britain's employment tribunal system is absorbing a surge of claims written with consumer chatbots rather than lawyers, according to a report published Sunday drawing on an internal memo from tribunal presidents Barry Clarke and Susan Walker.
Claims filed by workers against employers rose 39 percent to roughly 50,000 in the year to March 2026. Unresolved single cases climbed 55 percent over the same period, from about 45,000 to 64,000. The sharpest signal is procedural: applications for interim relief, once running at roughly 20 a year across all of Great Britain, are now arriving at about 20 a month, most of them attached to protected-disclosure claims.
Judges report that many of the longest submissions bear the marks of generative drafting — filings stretching to hundreds of pages, bundling dozens of separate legal claims at once, and in some cases resting on statutes that do not exist. Presidential guidance issued jointly by the two tribunal presidents took effect on 22 June, aimed at deterring speculative interim relief applications.
The economics underneath
The mechanism is a cost shift rather than a change in the law. Drafting a tribunal claim previously required either a solicitor's fee or substantial personal effort; a chatbot removes both barriers, while the tribunal system charges claimants no fee to file. The cost of a marginal claim falls close to zero for the person bringing it, but every filing still consumes judicial time and forces employers to fund a defence — valid and speculative claims alike. The report characterises the result as a tragedy of the commons in an AI edition.
Pressure is set to increase. Labour's Employment Rights Act introduces roughly 25 additional grounds on which a claim can be brought and removes compensation caps in some categories.
Why it matters
This is among the first hard data showing generative AI restructuring the economics of an entire public institution rather than a single workplace. The dynamic is not British: the same reporting notes comparable strain appearing in US federal courts, and any jurisdiction with low-cost filing and a self-representation tradition faces the same arithmetic. It also sharpens a question regulators have so far treated abstractly — whether consumer AI products carry any responsibility for the volume, and the accuracy, of the legal documents they produce.