UK judiciary updates AI guidance

Reading time: two minutes

The Courts and Tribunals Judiciary has published an updated version of its AI guidance for judicial office holders, clerks, judicial assistants and other staff. The new guidelines outline the key risks and issues associated with using AI and includes an expanded glossary of common AI terms, such as AI prompt, generative AI chatbot, hallucination and responsible AI.

The guidance states that, while AI tools are potentially helpful for summarising large bodies of text, writing presentations and performing administrative tasks, users should understand potential drawbacks before using AI tools in courts and tribunals.

For example, the guidelines emphasise that public AI chatbots don’t provide answers from authoritative databases. Instead, they generate text based on algorithms and the data they have been trained on, which may be “inaccurate, incomplete, misleading or out of date”. In addition, the framework stated that AI tools aren’t a reliable way to research new information that can’t be verified, and explained that the quality of AI’s answers depends on the nature of the prompts and the strength of the underlying data sets.

Given these limitations, the guidance says the output “must be checked before it is used or relied upon” and doesn’t recommend using AI for tasks such as legal research and analysis. The document highlighted: “Any use of AI by or on behalf of the judiciary must be consistent with the judiciary’s overarching obligation to protect the integrity of the administration of justice.”

Another section cites giveaways that work may have been produced using AI, including:

• references to unfamiliar cases or cases with unfamiliar citations (sometimes from the US);
• parties citing different bodies of case law in relation to the same legal issues;
• submissions that don’t accord with the general understanding of the law in the area;
• submissions that use US spelling or refer to overseas cases; and
• content that appears to be highly persuasive and well written but contains an obvious substantive error upon closer inspection.

The Lady Chief Justice of England and Wales, Baroness Carr of Walton-on-the-Hill, endorsed the updated guidance, along with other senior judicial figures. She said: “The growing accessibility and relevance of AI in the court and tribunal system means it is important that its use by or on behalf of the judiciary is consistent with its overarching obligation to protect the integrity of the administration of justice.”