Large language models have led to a slew of pro se lawsuits where people believe they can represent themselves with the help of AI. You’d expect better for actual lawyers, but it turns out you’d be wrong. As 404 Media first reported, a Mississippi judge recently tossed a case and sanctioned lawyers on both sides for submitting court filings riddled with AI-fabricated information.
The case was a dispute between lawyer Tom Withers and the city of Aberdeen, Mississippi, in which Withers alleged that he was owed unpaid legal fees. Withers wasn’t representing himself, but maybe in retrospect wishes that he had, because his representation decided to use AI to conduct research for the case and to actually generate the filings submitted to the judge. Per a court document, the two lawyers representing Withers admitted that they didn’t verify any of the information produced by AI before filing their briefs.
That’s not great, but it gets worse, because the lawyers representing the city of Aberdeen also used AI, meaning professional lawyers were basically wasting everyone’s time by making LLMs argue against each other.
As you might imagine, the judge in the case was not exactly thrilled by the whole situation. Sharion Aycock, senior United States District Judge for the Northern District of Mississippi, took the dramatic step of pausing the proceedings entirely and canceling the trial for the time being while dismissing all four lawyers involved in the case. Two of those lawyers—one from each side, both of whom admitted to using AI tools to generate their filings—are barred from appearing before the court for two years. All lawyers involved also got hit with fines, ranging from $1,000 to $3,500, based on whether they simply failed to verify information in a filing or if they actually drafted documents with AI-hallucinated citations.
“This case presents the Court with an unusual scenario—attorneys for both litigants engaged in similar sanctionable conduct,” Aycock wrote in a sanctions order. “This court is yet again ‘burdened with addressing AI hallucinations court filings.’”
The AI problem has hit courts across the country pretty hard, much to the dismay of judges who find themselves tasked with checking the work of lawyers to make sure some completely hallucinated citation doesn’t accidentally get included in actual legal precedent. Law researcher Damien Charlotin has taken it upon himself to track every instance of AI-generated citations in legal filings, and has so far documented a jarring 1598 cases so far. If you are a lawyer, please help Charlotin by simply not using AI slop in your work. It’s the least you can do.
Read the full article here
