In Fletcher v. Experian Information Solutions, Inc., No. 25-20086 (5th Cir. Feb. 18, 2026), the court — in an opinion by Chief Judge Elrod, joined by Judges Smith and Wilson — sanctioned counsel $2,500 after concluding that she used generative AI “to draft a substantial portion, if not all, of her reply brief”; failed to verify the brief's hallucinated quotations and citations; and then offered evasive, misleading explanations when the court asked what happened. 

Two aspects of the opinion stand out. First, the court reaffirmed that a bespoke rule is unnecessary to govern AI misuse in lawyers' briefing. Existing tools — Federal Rule of Appellate Procedure 46(c) (allowing courts to discipline attorneys for “conduct unbecoming a member of the bar”) and the court’s inherent authority (allowing the imposition of "sanctions for abuse of the judicial process”) — are enough to address inaccurate filings, as well as an attorney's misleading behavior when confronted with such errors. The Fifth Circuit even declined last year to adopt a proposed certification rule about AI use, opting instead to remind practitioners that they must review filings for accuracy. 

Second, candor matters. The court emphasized that it likely would have imposed lesser sanctions had counsel been forthright about the mistake from the outset. Instead, after the show‑cause order itemized a laundry list of fabricated quotations and other misstatements, counsel blamed “publicly available sources” for the errors — which, as the court showed, was highly unlikely (to put it mildly) — and only later confessed to the use of AI. The court deemed this conduct “evasive, misleading, and sanctionable.” 

Takeaways: Use AI responsibly — not blindly. (Or, as the maxim goes, “Trust, but verify.”) And if you discover an error after filing, notify the court and own up to it — promptly. The Fifth Circuit’s message isn't “don’t use AI”; it’s “don’t outsource your professional judgment — and take accountability when you slip.”