Summary
In *Fletcher v. Experian Information Solutions, Inc.*, the Fifth Circuit sanctioned attorney David N. Myers for filing a reply brief prepared in substantial part with AI that contained fabricated quotations and other inaccuracies. The opinion is especially useful because the court did more than punish the mistake: it explained why existing appellate rules already cover AI misuse and noted that the failure pattern is shifting from fake cases to false quotations drawn from real ones.
Why It Matters
This is a high-value direct legal story because it shows exactly how AI misuse is being operationalized in appellate practice:
- lawyers remain personally responsible for every quotation and proposition in AI-assisted briefing
- courts may treat a lawyer's lack of candor after the error is found as a separate aggravating factor
- existing signature and factual-certification rules can do the work even without an AI-specific local rule
- hallucination risk now extends beyond invented authorities into distorted excerpts from genuine authorities
The most important signal is doctrinal: the court rejected the idea that a new AI rule was necessary before sanctions could attach.
What the Source Says
The Fifth Circuit found that Myers used AI to draft a substantial portion of the reply brief and failed to check the quoted material carefully before filing. The court issued a $2,500 sanction after concluding that the brief contained inaccurate quotations from real cases and that Myers was not forthcoming when asked to explain the errors. The opinion also discussed a proposed local Rule 32.3 focused on AI and declined to adopt it, reasoning that existing Rule 32.2, Federal Rule of Appellate Procedure 32, and counsel's signature obligations already prohibit inaccurate factual or legal assertions regardless of whether AI generated them.