Anonymous Plaintiff had previously attempted to sue pseudonymously. Another 2024 SDNY case.
After Second Circuit appeal dismissed, new state suit removed to SDNY; remand denied and Plaintiff ordered to refile in her name or face dismissal.
Anonymous v. New York City Dept. of Education, et al., (S.D. NY 2024): Overview
Plaintiff cited cases that the Defendants and Court were unable to locate (no specific examples provided). [Source]
Court did cite Mata v. Avianca (S.D. New York) and Park v. Kim (2nd Cir. 2024) and the Court speculated that artificial intelligence (such as ChatGPT) may have been used, but basis was nonexistent case citations.
You can find a map with prior blog posts on cases here.
Outcomes:
The Court denied Plaintiff’s motion to remand and two other motions.
Plaintiff required to file a new version of her complaint using her current legal name (no pseudonyms or anonymity) or the Court would dismiss the action.
Mentioned potential sanctions in citing Park v. Kim and Mata v. Avianca, but did not explicitly cite FRCP Rule 11.
Interesting Quotes
“No secret” that LLMs hallucinate
V. Nonexistent Legal Authority
One last matter remains for the Court to address. Defendants note that, at times, Plaintiff “cit[es] to and reli[es] on what appears to be non-existent legal authority. Remand Opp. at 9. Having reviewed the case citations flagged by Defendants, the Court is likewise unable to locate them. Without question, it is improper and unacceptable for litigants –including pro se litigants – to submit “non-existent judicial opinions with fake quotes and citations.” Mata v. Avianca, Inc., 678 F. Supp. 3d 443, 448 (S.D.N.Y. 2023); see Tragath v. Zuck, 710 F.2d 90, 95 (2d Cir. 1983) (the right to self-representation “does not exempt a party from compliance with relevant rules of procedural and substantive law” (citation omitted)).
Also, assuming that (as was true in Mata) these nonexistent cases are the product of Plaintiff using an artificial-intelligence program like ChatGPT, see 678 F. Supp. 3d at 451, it is no secret that such programs can be unreliable, see, e.g., Cade Metz, Chatbots May “Hallucinate” More Often Than Many Realize, N.Y. Times (Nov. 6, 2023), https://www.nytimes.com/2023/11/06/technology/chatbots-hallucination-rates.html (“[One study] estimates that even in situations designed to prevent it from happening, chatbots invent information at least 3 percent of the time – and as high as 27 percent.”). Anonymous v. NYC Department of Education, et al. (SDNY 2024).
Takeaway: Ignorance of AI hallucinations is not an excuse. The Court noted—in 2024 and for a pro se litigant—that “it is not secret that such programs can be unreliable,” referencing a New York Times article as an example. Yet recently, in a Northern District of Mississippi case, four attorneys were disciplined because of hallucinated citations, with one attorney stating she was “unaware that A.I. could produce hallucinated cases and explained that she did not even know what a hallucinated case was” (see below).
Compare to N.D. Mississippi: “For these reasons, the Court disbelieves [] claim of ignorance…”
In fact, the Fifth Circuit has recently held that “[i]f it were ever an excuse to plead ignorance of the risks of using generative AI to draft a brief without verifying its output, it is certainly no longer so.” Fletcher, 168 F.4th at 235. For these reasons, the Court disbelieves Wilson’s claim of ignorance of the risks associated with using AI in the legal context. The Court finds that she knew, or reasonably should have known, of those risks and that she acted in bad faith in failing to verify the legal authority in her AI-generated brief. See In re Sealed Appellant, 194 F.3d at 671 (holding that “[w]hen bad faith is patent from the record and specific findings are unnecessary to understand the misconduct giving rise to the sanction, the necessary finding of ‘bad faith’ may be inferred.”). Withers v. City of Aberdeen (N.D. Miss. 2026)
Takeaway: It’s 2026. Attorneys need to understand how generative AI works if they are using it. And for attorneys who think they do not use AI, like the local counsel in this case, attorneys should understand how other people they work with are using AI (e.g., local and pro hac co-counsel, paralegals, interns, junior associates, etc.).
Compare to N.D. Illinois: Ignorance of AI hallucinations “without checking the results is in some sense irrelevant.”
The first reason I issue sanctions stems from [attorney’s] claim of ignorance—he asserts he didn’t know the use of AI in general and ChatGPT in particular could result in citations to fake cases. (Dkt. No. 71 at 3) [attorney] disputes the court’s statement in Wadsworth [Wadsworth v. Walmart (D. Wyoming 2025)] that it is “well-known in the legal community that AI resources generate fake cases.” 348 F.R.D. at 497.
[…]
This has been a hot topic in the legal profession since at least 2023, exemplified by the fact that Chief Justice John G. Roberts, Jr. devoted his 2023 annual Year-End Report on the Federal Judiciary (in which he “speak[s] to a major issue relevant to the whole federal court system,” Report at 2) to the risks of using AI in the legal profession, including hallucinated case citations. To put it mildly, “[t]he use of non-existent case citations and fake legal authority generated by artificial intelligence programs has been the topic of many published legal opinions and scholarly articles as of late.”
[…]
Counsel’s professed ignorance of the dangers of using ChatGPT for legal research without checking the results is in some sense irrelevant. Lawyers have ethical obligations not only to review whatever cases they cite (regardless of where they pulled them from), but to understand developments in technology germane to their practice. [Footnote 11: See, e.g., ABA Model Rule 1.1, Comment 8, made applicable here by Local Bankruptcy Rule 9029-4A (and applicable to all Illinois lawyers following adoption by the Supreme Court of Illinois), requires lawyers to “keep abreast of changes in the law and its practice, including benefits and risks associated with relevant technology.”] In re Marla C. Martin (N.D. Ill. Bankr. 2025)
Takeaway: I cover this case and other specific examples in detail in my Ethics CLE course, “AI Gone Wrong in the Midwest.”
