'This case presents the Court with an unusual scenario': Judge kicks lawyers off cases are finding out both were using AI to argue

Mississippi judge takes a dim view of proceedings once it was clear lawyers on both sides were using AI to make their arguments
The sanctions order included fines, disbarment, and disqualification from the case at hand for the lawyers involved
Advice from AI remains a somewhat tricky affair, given the lack of accountability and the model’s tendency to ‘hallucinate’

In what can be considered a comical occurrence that could be a sign of things to come, a US federal judge had to manually step in and admonish lawyers on both sides of the aisle after she noticed they were using AI.

Senior US District Judge Sharion Aycock, of the Northern District of Mississippi, noted that this was not the first time her court had to deal with the matter of being “burdened with addressing AI hallucinations in court filings.”

The judge ordered a pause in proceedings, scrapping the trial, while disqualifying all four lawyers involved from the case at hand and barring two of them from appearing in any case in the local Northern District of Mississippi for two years.

An ‘insufficient and incredulous’ justification

While the case might have been a routine one in the judge’s docket, addressing a breach-of-contract claim over unpaid legal fees between the city of Aberdeen and a Louisiana attorney, Tom Withers III, some of the precedents cited in the argument never occurred, inviting the judge’s scrutiny and eventual

His lawyer, Kathleen M. Wilson, used AI-hallucinated citations to argue their position, a situation that was uncovered when a court-mandated order required both sides to produce copies of the cases they had cited.

The city of Aberdeen, represented by Kathryn Y. Williams, was also found culpable for a similar offense: citing a non-existent 1971 Mississippi Supreme Court decision and references to three other federal rulings that could not be reproduced.

Both lawyers admitted to using AI while claiming ignorance of the potential of the LLMs they employed, at times hallucinating. The judge, however, took a dim view of the entire affair, noting that one of the lawyers had been practicing for at least six months using generative AI to draft her cases without oversight, and that they had previously been warned against the practice in an unrelated case.

The judge noted that she “finds that explanation to be insufficient and incredulous,” and fined the four lawyers a total of $8,000, singling out the two lawyers who used AI.

The case does, however, mark an important ruling that could, ironically, serve as a real precedent against the AI-generated ones that got the lawyers into trouble: ignorance of AI’s hallucinations is not a viable legal defense.