Judge Disqualifies Lawyers Over AI-Hallucinated Citations

U.S. District Judge Sharion Aycock issued a sanctions order on June 8 after finding that multiple court filings in a contract-fee dispute contained hallucinatory, AI-generated legal citations, according to Reuters, Bloomberg Law and the ABA Journal. The judge disqualified all four attorneys on the two sides of the case and revoked the pro hac vice admissions of the two out-of-state counsel, per Bloomberg Law and ABA Journal. Aycock also barred the two out-of-state lawyers from appearing in the Northern District of Mississippi for two years and imposed monetary sanctions totaling $8,000; Reuters reports the fines as $2,500 and $3,500 for the out-of-state lawyers and $1,000 each for the local counsel. The dispute concerns a contract claim by Tom Withers III against the City of Aberdeen, per Bloomberg Law.

Large language models and many generative-retrieval setups can fabricate or misattribute case names, citations and statutory text when prompted for legal research. Reported filings in this matter contained authorities the court could not verify, a pattern regulators and technologists have documented when LLMs are used for legal-research tasks without authoritative, citation-anchored retrieval layers. The central professional-liability issue is the distinction between surface-level LLM drafting and verified legal research workflows that confirm citations against primary sources or commercial databases before filing.

Courts and bar regulators have increasingly confronted instances where lawyers used AI tools to draft pleadings or research and submitted unverified material to judges. Bloomberg Law notes that Aycock in a separate matter previously imposed more than $20,000 in sanctions and required continuing legal education on AI hallucinations; Reuters and the ABA Journal report multiple recent cases where judges sanctioned attorneys for similar conduct. The Mississippi order was unusually broad because it disqualified counsel on both sides and revoked pro hac vice admissions, signaling judicial intolerance for filings containing fabricated authorities in litigated matters.

For practitioners: Monitor state bar guidance, district-court pilot programs and continuing-legal-education requirements related to AI. Bloomberg Law reports the court sent the order to state bar associations and required CLE for at least one sanctioned attorney. Observers should also watch for updated local rules or court guidance that prescribe verification steps for AI-assisted research, and for law firms to adopt explicit citation-verification checkpoints before filing documents.

For practitioners: The sanctions reinforce that courts treat the accuracy of cited authorities as an attorney responsibility, regardless of the tools used to prepare filings. Law firms and legal teams will likely need documented verification processes for any research or drafting that relies on generative systems to avoid professional-discipline and litigation risks.