Judges Say AI Can't Replace Human Judgment In Courts

While artificial intelligence holds promise for improving court efficiency, its current limitations, including inaccuracies and ethical concerns, make human judgment indispensable in legal proceedings, a group of judges said during a panel discussion at a recent legal technology conference.

The final day of the 2025 Everlaw Summit in San Francisco on Thursday featured a session where judges shared their practical experiences on digital tools, discussing what works and doesn't work in the courts. Gloria Lee, Everlaw's chief legal officer, moderated the panel of judges from all around the country.

"The courts are really behind the times," U.S. District Judge Joy Conti of the Western District of Pennsylvania said during the panel session.

Judge Conti said the courts are being cautious with the use of AI, because the tech is not reliable today due to its ability to produce inaccurate information.

Alameda County Superior Court Judge Victoria Kolakowski said during the panel that the courts are moving slowly due to data security issues with AI that make it difficult to use for confidential information.

Judge Conti said the courts are at least four to five years away from the expanded use of AI in the courts, saying that AI cannot see the nuances of the facts in a case now, and it's better to have human judges read filings.

"There is nothing like reading the case yourself," Judge Conti said. "There is nothing that can substitute for the lawyer and the judge actually reading the cases."

One issue plaguing courts is the submission of so-called deepfakes in evidence, which are AI-generated images and videos that look real. This year, Judge Kolakowski ruled over a case in which pro se litigants submitted deepfake evidence and argued that it was just an enhancement of the facts.

"What's a deepfake and what's an enhancement?" Judge Kolakowski said. "What's real and what's not?"

Los Angeles County Superior Court Judge David Cunningham said during the panel that, while AI tools can create efficiency, the human element is "indispensable," because otherwise you might get a wrong answer.

"You can't get rid of the human element," Judge Cunningham said.

Most of the judges in the panel agreed that AI could be beneficial in the long run.

Chief Magistrate Judge Willie Epps Jr. of the Western District of Missouri said during the panel that the next few years will be tough, as some users misuse AI, but the technology has the opportunity to help litigants who may not be able to afford a lawyer. He expects the tools to change things and help users save time.

"You are all changing the world," Judge Epps said to the legal tech crowd in attendance.

Judge Cunningham said the filings in Los Angeles County court have doubled since the pandemic, and that AI could be used to cut that down. He also said AI can be useful in e-discovery and mass torts, helping lawyers get organized during potential discovery disputes.

U.S. Magistrate Judge Young B. Kim of the Northern District of Illinois said during the panel that he uses AI tools for summarizing long opinions, creating timelines and drafting small paragraphs on issues that he can easily verify. However, he also said that users need to make sure that the output of these tools are accurate.

As more AI tools make their way into the courts, Judge Kim said that perception is still key, and the judicial branch needs to come across as fair.

"Efficiency is important, but the perception of being fair is equally important," Judge Kim said.

--Editing by Nicole Bleier.

For a reprint of this article, please contact reprints@law360.com.