Federal judges admit staff used AI for error-ridden court orders

Two federal judges admitted that members of their staff used artificial intelligence to prepare court orders over the summer that contained errors.

The admissions, which came from U.S. District Judge Julian Xavier Niles in New Jersey and U.S. District Judge Henry Wingate in Mississippi, came in response to an inquiry by Sen. Chuck Grassley, R-Iowa, who chairs the Senate Judiciary Committee.

Grassley called the court’s recent orders “full of errors.”

In letters released by Grassley’s office on Thursday, the justices said the rulings in the cases, which were not related, had not undergone their chambers’ usual review processes before they were released.

Trump administration unveils groundbreaking tool to ‘supercharge’ government efficiency to ‘win the race’ for AI dominance

The two judges said they had since adopted procedures to improve how rulings are reviewed before they are published.

A draft decision dated June 30 in a securities class action lawsuit “was released in error — human error — and was withdrawn as soon as it was brought to my chamber’s attention,” Niles said in his letter.

The judge said a law school intern used OpenAI’s ChatGPT to conduct legal research without authorization or disclosure, which he also said was contrary to chamber policy and relevant law school policy.

“My chamber policy prohibits the use of GenAI in legal research or drafting opinions or orders,” Niles wrote. “In the past, my policy was communicated verbally to chamber staff, including interns. This is no longer the case. I now have an unambiguous written policy that applies to all law clerks and interns.”

Federal judge imposes fines and reprimands lawyer who used AI to draft court filings

Senator Chuck Grassley called the court’s recent orders “full of errors.” (Tom Williams/CQ-Roll Call, Inc via Getty Images)

A law clerk used Perplexity “as a foundational drafting aid to compile publicly available information on the list,” Wingate said in his letter, adding that publishing the draft decision on July 20 “was an error in human oversight.”

“This was a mistake. I have taken steps in my chambers to ensure this mistake will not be repeated again,” the judge wrote.

Wingate had removed and replaced the original order in the civil rights lawsuit, declining at the time to provide an explanation but saying it contained “clerical errors.”

Grassley had asked the justices to clarify whether artificial intelligence had been used in the decisions after lawyers in the cases in question raised concerns about factual inaccuracies and other serious errors.

Clear errors in artificial intelligence force two judges to withdraw separate rulings

Senator Chuck Grassley asked the justices to clarify whether artificial intelligence was used in the decisions after lawyers raised concerns about factual inaccuracies and other errors. (Photo by Susan Walsh/Pool/AFP via Getty Images)

Click here to download the FOX NEWS app

“Honesty is always the best policy. I applaud Judges Wingate and Niles for admitting their mistakes and I’m glad to hear they are working to make sure this never happens again,” Grassley said in a statement.

The senator continued: “Every federal judge, and the judiciary as an institution, has an obligation to ensure that the use of generative artificial intelligence does not violate the rights of litigants or prevent fair treatment under the law.” “The judiciary needs to develop more decisive, purposeful, and durable AI policies and guidelines. We cannot allow laziness, apathy, or over-reliance on artificial assistance to undermine the judiciary’s commitment to integrity and factual accuracy. As always, my stewardship will continue.”

Lawyers have also faced scrutiny from judges across the country over accusations of misusing artificial intelligence in court filings. In response, judges have issued fines or other penalties in several cases over the past few years.

Reuters contributed to this report.