Federal judges increasingly use AI to draft rulings and prepare for hearings, study finds

A Northwestern University study indicates a growing trend of federal judges incorporating Artificial Intelligence into their workflow. Over 60% of judges are now using AI tools for tasks ranging from analyzing case filings and preparing for hearings to drafting rulings and identifying legal arguments, with about 22% using them daily or weekly. Judges like Xavier Rodriguez report significant time savings, using AI to quickly generate case timelines. Courts are formalizing this shift through pilot programs, such as the one launched by the Los Angeles County Superior Court with Learned Hand, and contracts with companies like Thomson Reuters and LexisNexis.

Despite the benefits of increased speed and efficiency, concerns remain regarding the reliability of these AI tools. A Stanford University study found error rates between 17% and 33% in legal AI queries, and instances of 'hallucinations' – the generation of nonexistent cases or false quotes – have already occurred in court filings. Judges who experienced these errors attributed them to clerks and interns using AI, with one judge, Julien Xavier Neals, banning generative AI from his chambers.

Crucially, judges emphasize that AI serves as a drafting aid and research assistant, not a decision-maker. They maintain final authority and stress the importance of independent judgment. While adoption is increasing, it remains in its early stages, with courts proceeding cautiously through pilot programs and careful testing. Vendors are working to improve reliability by sourcing information from verified legal databases and providing links to source cases.

(Source：Complete Ai Training)