Jimenez-Fogarty v. Fogarty (S.D.N.Y.): a lawyer who would not say where her fake citations came from is fined $2,500 under Rule 11 for AI hallucinations in bad faith

The problem

Tricia S. Lindsay, attorney for plaintiff Sai Malena Jimenez-Fogarty, “responded to two motions to dismiss by filing a pair of memoranda of law that cited to numerous nonexistent cases” [source]. The court counted as fabricated only citations it could not find at all, or that pointed somewhere unrelated, and set aside typos and wrong case names ([source]). One example: the brief cited “Cornhill LLC v. Sowers, 183 A.D.3d 649” as upholding nail and mail service, and the court wrote, “We cannot find any case by this name” [source]. Ordered to describe exactly how the briefs were drafted, Lindsay acknowledged that eleven citations were inaccurate but, in the court’s words, “she provides no explanation of how these citations came to be included in the memoranda of law” [source].

What was built

Nothing was built. This is an honest-negative: an AI tool’s output reached a federal court unchecked, and the lawyer would not say which tool. The court reasoned that fabricated citations come from a person or from an AI program, noted Lindsay denied any other person was involved, and held: “We thus find that Lindsay used an AI program that generated the fabricated citations in the memoranda of law and that Lindsay failed to check that these citations were genuine” [source]. Lindsay had suggested LexisNexis’s AI features might be the source; the court concluded the citations came from “some kind of AI program (though we do not credit that it was any such program supplied by Lexis Nexis, as Lindsay at one point hints)” ([source]). So this record pins no product, and this page names none.

The outcome

The court applied the rule that courts in the Second Circuit “have repeatedly found that presenting AI-generated hallucinations as valid caselaw constitutes subjective bad faith,” and found clear and convincing evidence of bad faith here [source]. It pointed to repetition after the orders to show cause: “she filed a brief in the United States Court of Appeals for the Second Circuit that contained two fabricated citations,” and of the eight citations challenged in a later objection it concluded “they are in some cases completely fabricated” [source]. The order reads: “Tricia S. Lindsay is sanctioned pursuant to Rule 11 of the Federal Rules of Civil Procedure in the amount of $2,500.00,” payable to the Clerk by May 14, 2026 [source]. It declined to award the defense attorneys’ fees ([source]). Because “it is important that other courts be aware of Linsday’s repeated misconduct,” it ordered her to alert the judge in each pending case where she appears as attorney of record, and to give her client a copy [source]. For scale, the opinion cites a public tracker identifying “916 proceedings in U.S. courts to date involving citations to non-existent cases likely ‘hallucinated’ by AI, including 342 where lawyers supplied the offending citations”; that is the court quoting a tracker, not a TIN count [source].

The ruling is not final on the record read. Lindsay filed an “Objection to Report and Recommendations” on May 28, 2026 [source], and the order itself says a timely objection extends its deadlines until the objection is decided ([source]). No ruling on it appeared on the docket as of October 31, 2026 [source].

Other courts already rely on the holding. In a September 14, 2026 Rule 11 order, Judge Cathy Seibel cited Jimenez-Fogarty, reported at 831 F. Supp. 3d 253, for the rule that “reliance on nonexistent legal authority fabricated by GenAI constitutes bad faith under Rule 11” [source]. The Volokh Conspiracy reported the sanction on May 2, 2026, noting that Lindsay had run for the New York State Senate in 2024 ([source]). On May 20, 2026, Buffalo litigators Dan D. Kohane and Michael J. Williams of Hurwitz Fine P.C. analysed the opinion in a bylined article and wrote that “the court ordered a $2,500 fine” [source].

The weakest load-bearing source, named where you meet it

The $2,500 figure rests on the signed opinion, a Tier 1 court record. Its only independent press corroboration is The Volokh Conspiracy (Tier 2), and that post is mostly an excerpt of the opinion with a short bylined introduction, so it confirms the opinion says what this page says but adds no separate reporting on the amount [source]. The second independent write-up of the amount is attorney commentary in a law-firm publication (Hurwitz Fine P.C., Tier 2), not a newsroom; its authors and firm do not appear as counsel on the docket, its $2,500 matches the opinion exactly, but its paraphrase of the notice orders is looser than the opinion’s own wording, so this page states those orders from the opinion only ([source]). The AI and bad-faith finding has a stronger second reader: a different S.D.N.Y. judge’s order that cites it [source]. The finality point rests on a docket read through CourtListener that could not be archived this session; a local copy is kept instead ([source]).

Conflation guards

• The $2,500.00 is the only sanction in this order. The court considered and declined a fee award to defendant Emily Hirshowitz.
• No AI product is pinned. The court found “an AI program” and expressly declined to credit LexisNexis as the source; counsel never named a tool.
• The 916 and 342 figures are the court citing Damien Charlotin’s tracker, not findings about this case and not a TIN count.
• This is a magistrate judge’s ruling under objection, not a district judge’s affirmance.

Related case files

• The canonical S.D.N.Y. AI-hallucination Rule 11 sanction in Mata v. Avianca: the same court and the same rule, and the opinion this one cites for the subjective bad-faith standard, three years earlier.
• Flycatcher v. Affable Avenue, where repeated AI-fabricated citations cost a client the case by default judgment: the far end of the remedy range this opinion surveys, from the same district in the same year.
• A federal appeals court fines a lawyer $2,500 for an AI-drafted brief in Fletcher v. Experian: the same dollar amount, and the same lesson that a lawyer who is not forthcoming about AI use draws a heavier response.