Robot Brief Faceplant

The 9th Circuit sanctioned lawyers over AI-hallucinated briefs, because apparently fake cases are still bad even when autocomplete wears a tie

The 9th Circuit fined two California lawyers $2,500 each and suspended them from practicing before the appeals court for six months over briefs with nonexistent cases and misrepresented authorities.

What Happened

The 9th U.S. Circuit Court of Appeals sanctioned California lawyers Mike Singh Sethi and William Rounds after finding they filed briefs containing multiple nonexistent cases, misattributed quotations and gross misrepresentations of real cases.

ABA Journal reported that the lawyers initially claimed the errors were typographical mistakes and repeatedly denied that generative AI had produced the problems. The court said the lawyers later admitted it was "probable" the errors came from unauthorized AI use by brief writers and that they had not checked the citations before filing.

The court imposed a $2,500 sanction on each lawyer, suspended them from practicing before the 9th Circuit for six months and required them to disclose any AI use in filings for the next two years. The Los Angeles Times noted the underlying immigration clients had already won a separate ruling halting deportation, and the disciplinary opinion did not undo that result.

Why This Matters

This is not a "technology is scary" story. Courts do not care whether a bad citation came from a chatbot, a sleep-deprived associate or a cursed spreadsheet. Lawyers sign filings. Lawyers own the words. Lawyers are supposed to read the cases they cite.

The especially important part is candor. The 9th Circuit said lesser sanctions may have been warranted if the lawyers had promptly disclosed the AI issue and apologized. Instead, the court said the misconduct continued through a motion to correct, oral argument, an order-to-show-cause response and other filings.

The Dumb Part With The Fake Precedent

The dumb part is pretending fake case law is a formatting issue. A nonexistent citation is not a typo. It is a legal scarecrow wearing a Westlaw hat.

Generative AI can be useful for drafts, summaries and boring first-pass work. But the moment it invents a court case, the adult in the room has to stop, verify and tell the truth. "The robot hallucinated and we did not check" is embarrassing. "No, Your Honor, that was probably just a typo" is how embarrassment becomes a published opinion.

The Bottom Line

The 9th Circuit's warning is blunt: read everything cited in a court filing, be aware of overreliance on generative AI and disclose AI hallucinations quickly and transparently. The real stupid shit is that the legal profession needed another reminder that imaginary cases do not become law just because a chatbot says them with confidence.

Sources

U.S. Court of Appeals for the 9th Circuit: June 3, 2026 disciplinary opinion

ABA Journal: Attorneys' lack of candor over AI errors leads to stricter sanctions

Los Angeles Times: Suspended O.C. immigration attorneys filed briefs filled with AI slop


← Back to Internet Nonsense