UK court warns lawyers of severe penalties for misuse of AI

In a significant development, the High Court of England and Wales has issued a strong warning to legal professionals about the dangers of using generative artificial intelligence (AI) tools like ChatGPT for legal research without proper verification.

In a ruling combining two recent cases, Judge Victoria Sharp emphasised that while AI can produce convincing responses, these can often be inaccurate or entirely false.

“Such tools can produce coherent and plausible responses to prompts, but those responses may turn out to be entirely incorrect,” Judge Sharp wrote.

She reiterated that lawyers have a professional duty to ensure the accuracy of any legal research by consulting authoritative sources, especially when using AI-generated content in court submissions.

It should be noted that the ruling follows two troubling cases. In one, a lawyer representing a man suing two banks submitted 45 legal citations, 18 of which were completely fabricated. 

Many others were either misquoted or irrelevant to the matter at hand.

In another case, a lawyer defending a man evicted from his London home cited five non-existent legal precedents.

Although the lawyer claimed not to have used AI directly, she acknowledged that AI-generated summaries from search engines may have influenced her references.

Judge Sharp also stressed that while the court chose not to initiate contempt proceedings in these instances, this should not be viewed as setting a precedent.

She stated clearly: “Lawyers who do not comply with their professional obligations in this respect risk severe sanction.”

Notably, the ruling has been forwarded to professional bodies such as the Bar Council and the Law Society to promote stricter adherence to legal standards.