Are you ready for the justice of the future? British judges allowed to write decisions using ChatGPT


Are you ready for the justice of the future? British judges allowed to write decisions using ChatGPT

December 14, 01:33 Share:

Judges in the UK will use ChatGPT to write court decisions (Photo: ramirezom/depositphotos)

The UK Courts Service has officially allowed judges to use ChatGPT and other artificial intelligence tools to write court decisions.

The regulator's decision opens up the opportunity for British judges to freely experiment with new technology. To avoid potential problems, the service has published a six-page guide for judges. It provides basic information about the technology and various caveats, providing guidance on how and for what tasks AI can be used.

“The use of artificial intelligence in society continues to grow, as does its relevance to the courts and tribunals system. This guidance is the first step in a proposed set of work to support the judiciary in its interactions with AI,” says the agency that oversees judges, magistrates and tribunals in England and Wales.

The manual, in particular, states that AI responses “may be inaccurate, incomplete, deceptive or biased”. ChatGPT and other large language models are not recommended to be used for research in order to find new information, and all responses received are strongly advised to be checked before being used.

The manual also warns that judges must be mindful of what data they enter as artificial intelligence companies collect results from users' interactions with chatbots.

Even before the publication of the manual, British judges managed to master ChatGPT and share their experience. In particular, in September, Lord Justice Birss from the Court of Appeal of England and Wales spoke at The Law Society conference and said that he uses ChatGPT in his work and considers this tool « very helpful». The chatbot, according to the judge, helped him summarize legal theories with which he was not familiar.

Read also: This is not funny anymore. ChatGPT was forced to reveal people's phone numbers and addresses using a simple attack. Dangerous precedent or trend? In Brazil, legislators adopted a document written by ChatGPT. Fake or really sick? ChatGPT got lazy, researchers rushed to diagnose “seasonal depression”