Most Federal Judges Have Used AI for Court Work, Study Finds

A majority of federal judges reported using an AI tool in their work and a third allow their staff to use the platforms, according to a new study.

The study from the New York City Bar Association and researchers with Northwestern University, publicized this week, found that over 60% of federal judges have used an artificial intelligence tool at least once in their judicial work. The platforms are most often used by judges and staff in their chambers for legal research.

But the study found that only 22.4% of judges use AI on a weekly or daily basis, and over a third have never used any of the tools they were surveyed about for work.

The paper was published in the Sedona Conference Journal, and US District Judge Xavier Rodriguez in San Antonio is a co-author.

The survey was sent to 500 randomly selected federal judges in December, and 112 responded. Bankruptcy judges had the highest response rate, while only six federal appellate judges submitted answers.

The surveyed bankruptcy and magistrate judges had the highest rates of using AI tools on a weekly or daily basis. Nearly half of district court judges said they “never” use AI tools in their work, according to the study.

While the vast majority of judges said their use of AI doesn’t touch their rulings, 1.8% surveyed said they use AI to “make decisions” and 4.5% said they use it to “inform decisions.”

The surveyed judges’ use of AI also highly correlated with the AI use of others in their chambers. But the study noted that other court staff were reported to use the tools more than the surveyed judges did, and that the judges might not know about all AI use in chambers.

Most judges—61.1%—said they hadn’t been offered AI training or didn’t know if it was offered. But for those who were offered it, 73.8% attended. “This suggests an unmet demand that more AI training should be offered and specifically in a visible manner,” the study reads.

And while roughly a third of judges said they allow AI use in their chambers, nearly a quarter of judges said they don’t have any official policy on AI use.

“The distribution of chambers policies—ranging from formal prohibitions to explicit encouragement within narrow guardrails—indicates that the judiciary is in an early phase of AI governance, with no dominant model yet emerging across chambers,” the study reads.

The federal judiciary last year created a task force to examine if the courts need new policies on AI.