Judges Cite AI-Generated Legal Citation Errors in Lawyer Sanctions

Two lawyers were barred after submitting legal filings with AI-generated fake citations.

Federal and state courts are increasingly encountering court filings containing fabricated legal citations created by generative AI tools. These so-called "hallucinations"—where AI invents cases or authorities—pose serious risks to legal proceedings and professional responsibility.

In a prominent case reported by The National Law Journal, attorneys Kathleen M. Wilson and Kathryn Y. Williams faced a two-year suspension from practicing in their district court. They were fined $2,500 and $3,500 respectively after submitting briefs containing numerous fabricated citations generated by AI without adequate verification. The court also sanctioned their local counsels for failing to supervise these errors, citing significant ethical breaches.

Retired U.S. Magistrate Judge Michael Wilner recounted an incident where he nearly included fake AI-generated citations in a judicial ruling, highlighting how such hallucinations can slip past even experienced judges, especially given the pressures and caseloads courts manage.

Research from Thomson Reuters Westlaw documented 22 occurrences of non-existent case citations identified and challenged in court filings during a 39-day period in late 2023. Legal researcher Damien Charlotin's ongoing study reported at least 95 total U.S. incidents involving AI hallucinated citations brought to light through mid-2023, underscoring an upward trend.

Some courts are enacting formal procedures. U.S. District Judge Brantley Starr implemented a requirement that any filings containing AI-generated content must include an attorney certification confirming human review and accuracy verification. Separately, the Illinois Supreme Court’s Attorney Registration and Disciplinary Commission published guidance reminding lawyers of their ethical duty to ensure AI-generated material submitted in court is truthful and verified.

While sanctions serve as deterrents, the evolving technology challenges courts to safeguard legal integrity. In a 2023 Kansas federal patent case, a judge fined $5,000 and revoked court admission for an attorney who submitted AI-hallucinated citations. These instances serve as a caution that unverified AI usage in legal filings can result in severe professional consequences.

For law firms and legal operations teams, these developments underscore the necessity of stringent review processes for AI-assisted document production. Careful integration of generative AI—with a focus on rigorous human verification—remains essential to maintaining ethical standards and procedural fairness in the courts.