Majority of US Federal Judges Now Use Artificial Intelligence, Study Shows

More than half of US federal judges — 60 per cent — are using at least one AI tool in their judicial work, according to a new study released on Monday.

Researchers from Northwestern University said theirs is believed to be the first random-sample study of AI usage among federal judges since the rise of AI in legal work following the 2022 release of OpenAI’s ChatGPT.

While 60 per cent of surveyed judges reported using AI at least occasionally for tasks such as reviewing documents, conducting legal research, and drafting or editing documents, only 22 per cent said they used the technology daily or weekly. Meanwhile, 38 per cent said they had never used AI in their judicial roles.

“Judges see the potential here,” said Northwestern law professor Daniel Linna, a co-author of the study alongside Northwestern computer science professor VS Subrahmanian, US District Judge Xavier Rodriguez, and Northwestern law student Siyu Tao. “Of course there are risks, but there are benefits if we get this right.”

The study comes amid growing concerns about AI-related errors in the legal field. Lawyers have faced court sanctions for citing “hallucinated” cases generated by AI, while two judges last year acknowledged that staff members had used AI to help prepare court orders that contained mistakes.

The new research is based on responses from 112 federal judges, drawn from a random sample of 502 federal bankruptcy, magistrate, district court, and appellate court judges — a sample large enough to offer insights into how federal judges are approaching AI, Linna said.

Judges were asked about their use of popular large language models such as ChatGPT and Claude, as well as AI tools designed specifically for legal use, including CoCounsel, Westlaw AI-Assisted, Deep Research, Protégé, Lexis+ AI, Vincent AI, Harvey, and Legora. The study found that judges were more likely to use legal-specific AI tools than general ones.

Among the respondents, 30 per cent said they used AI for legal research, making it the most common application. Reviewing documents was the next most cited use, at nearly 16 per cent.

One in three judges said they either permit or encourage AI use in their chambers, while 20 per cent formally prohibit it. Nearly 18 per cent discourage AI but do not prohibit it formally, and 24 per cent said they have no formal policy.

More than 45 per cent of judges surveyed reported that AI training was not provided by court administration — a finding that experts say could inform future policy decisions around court spending and training programs.

For any enquiries or information, contact ask@tlr.ae or call us on +971 52 644 3004. Follow The Law Reporters on WhatsApp Channels.