Ninth Circuit Suspends Two Lawyers for AI Hallucinations

Two California attorneys have been suspended from practice in federal appeals court for six months and fined $2,500 each after submitting legal briefs containing citations to nonexistent cases generated by artificial intelligence—and then denying they had used AI at all.

The U.S. Court of Appeals for the Ninth Circuit issued the disciplinary order Wednesday against Mike Singh Sethi and William Rounds, who practice together in Orange, California. The sanctions follow their submission of briefs in an immigration case that contained what the court called "gross misrepresentations of real cases."

The cover-up compounds the error

When judges questioned the fabricated citations, the attorneys initially blamed "innocent typographical mistakes" and denied using AI, according to the court's order. Only at an October hearing did Rounds acknowledge it was "possible" that a non-attorney staff member who drafted the briefs had relied on artificial intelligence.

The three-judge panel—comprising judges appointed by Presidents Clinton, Bush, and Trump—found that Sethi, who heads the firm, "violated his duties of competence and diligence." Both lawyers "violated their duty of candor" by mischaracterizing the errors as simple typos.

The court emphasized that while generative AI is not inherently unethical, "using generative AI without rigorously checking its output does present a higher risk of violating certain ethical and procedural rules." The judges warned that inaccurate case citations "may prove more dangerous to our profession in the long run" than outright fabrications.

Why it matters

This case illustrates the growing tension between efficiency gains from AI legal tools and professional responsibility standards. As law firms increasingly adopt AI to reduce research costs and speed up brief preparation, courts are establishing clear precedent that attorneys cannot delegate verification duties to non-lawyers or blame technology for their own failures. The Ninth Circuit's decision sends a signal that dishonesty about AI use will be punished more severely than the underlying technical errors.

Widespread problem in legal profession

The attorneys' misconduct reflects a broader pattern. One recent analysis identified 957 U.S. cases and nearly 1,400 worldwide involving nonexistent or falsely cited cases apparently produced by AI hallucinations. Multiple courts have now sanctioned attorneys for submitting fictitious case citations in legal arguments.

As part of their discipline, Sethi and Rounds must notify all clients and firm attorneys about the suspension. They must also declare under penalty of perjury in future filings whether AI was used in case preparation. The firm told the court it has hired a licensed attorney to review all briefs before submission.

Interestingly, the court did not withdraw its November ruling in the underlying immigration case, which granted the attorneys' request to halt deportation of three family members from India while their asylum claims are properly considered.

The San Francisco Chronicle first reported the disciplinary action. The attorneys did not immediately respond to requests for comment.