You know that sinking feeling when you realize you’ve been caught in a lie? Now imagine you’re a federal judge reviewing legal briefs for a straightforward contract dispute, only to discover that lawyers on both sides filed documents citing cases that don’t exist. Senior U.S. District Judge Sharion Aycock faced exactly this scenario in a Mississippi fee dispute between attorney Tom Withers and the city of Aberdeen.
According to 404 Media, all four attorneys involved either admitted using AI directly or submitted AI-prepared briefs without proper verification. The result? A canceled trial and sanctions that should make every lawyer using ChatGPT for research break into a cold sweat.
The Cascade of AI Failures
Attorneys continued using unreliable AI tools even after court flagged fabricated legal authorities.
The details read like a cautionary tale written by someone who’s never touched legal software. Lawyer Kathleen Wilson admitted using an AI tool called “First Drafts” to draft entire briefings, then—here’s the kicker—kept using AI even after Judge Aycock identified hallucinated cases in her filings.
Meanwhile, Kathryn Williams deployed an AI research tool designed for in-house work, not Mississippi law. This isn’t just about one bad prompt or a single oversight. Multiple attorneys, supposedly trained in legal research fundamentals, treated AI like Wikipedia circa 2005: authoritative until proven otherwise.
Courts Draw the Line
Unprecedented sanctions signal judicial patience with AI excuses has expired.
Judge Aycock didn’t just issue a stern warning. She canceled the trial, disqualified all four lawyers, and imposed fines between $1,000 and $3,500. Two attorneys face a two-year ban from appearing before her court. Her written order called the situation “unusual” because attorneys for both litigants engaged in similar sanctionable conduct.
This case joins a growing pattern of federal courts cracking down on AI hallucinations in legal filings, but the bilateral nature makes it unprecedented. When both sides of a legal dispute face identical professional responsibility violations, that’s not coincidence—that’s systemic failure.
The broader message cuts deeper than legal malpractice. Courts nationwide are establishing that “my AI did it” won’t fly as an excuse for filing fabricated authorities. Professional verification requirements aren’t suggestions—they’re the price of admission to practice law in an AI-enabled world.
