Mississippi Judge Bars 4 Attorneys Over AI Hallucinations in Court Fil
Senior U.S. District Judge Sharion Aycock of the U.S. District Court for the Northern District of Mississippi disqualified all four attorneys in *Withers v. City of Aberdeen* on June 8, 2026, after discovering that counsel for both the plaintiff and defendant had submitted court filings containing hallucinated legal citations—a watershed moment for AI accountability in federal litigation (FindLaw).
AI Tool Generated Four Fabricated Cases
The attorneys used "First Drafts," an AI drafting tool, to generate legal research. The tool produced four entirely fabricated case citations—references to judicial decisions that did not exist. When Judge Aycock's staff verified the citations, they discovered the hallucinations, prompting an immediate inquiry into counsel's conduct (Clarity Bot).
In her ruling, Judge Aycock stated: *"This case presents the Court with an unusual scenario—attorneys for both litigants engaged in similar sanctionable conduct."* The judge found that counsel "blindly" used AI in filings, resulting in "hallucinations, or fabricated quotes or sources"—a violation of Federal Rule 11 that requires attorneys to certify the accuracy of legal arguments (LawSnap).
Sanctions: Disqualification, Fines, and a Two-Year Ban
All four attorneys were removed from the case and their trial was canceled in June 2026. However, the consequences diverged by geography and standing:
Out-of-state attorneys: Wilson and Kathryn Y. Williams, both admitted *pro hac vice* (for this case only), were barred from appearing in the Northern District of Mississippi for two years. Wilson was fined $2,500; Williams fined $3,500 (FindLaw).
Local attorneys: The two resident attorneys admitted to the Northern District were disqualified from the case but not barred from the district. Each was fined $1,000.
The total sanctions reached $8,000. Judge Aycock also revoked both out-of-state attorneys' *pro hac vice* admissions and referred all four attorneys to their respective state bars for disciplinary action, a step that could trigger license suspension or revocation (PR Newswire).
Broader Implications for AI in Legal Practice
The ruling underscores a critical risk for law firms deploying large language models (LLMs) in document drafting without human verification. Hallucinations—AI-generated false information presented as fact—have been a known limitation of LLMs since their deployment in 2023. Yet the Withers case represents one of the first instances where a federal judge has imposed career-limiting sanctions (a multi-year ban from a district court) in response to AI-generated malpractice.
The case also demonstrates that AI risk is symmetric: both plaintiff and defense counsel fell into the same trap, suggesting systemic adoption of the tool without adequate guardrails. As courts begin to hold attorneys accountable for AI-enabled errors, the legal profession faces mounting pressure to implement verification protocols and disclosure requirements when using generative AI in filings.
