Generative artificial intelligence tools for lawyers ‘hallucinate’ up to one in three times

Generative artificial intelligence tools designed for the legal industry make up false or misleading information between 17 per cent and 33 per cent of the time, a Stanford University study has found.

The academic paper claims to be the first systematic assessment of leading AI tools for legal research tasks, which allow lawyers to interrogate large databases of legislation, case law and legal commentary by putting questions to an AI-powered chatbot.