Like Fraud and Spiderman, With Great AI Powers, Come Great Responsibilities …
During my weekly exercise of scouring the decisional databases for noteworthy and instructional fraud decisions, lo and behold, what do I find? A decision of the New York, Appellate Division, Third Department, that is admittedly unremarkable for its fraud-related merits disposition, but indeed noteworthy “as the first appellate-level case in New York addressing sanctions for the misuse of GenAI,” i.e., hallucinations on steroids.
The decision was rendered on January 8, 2026, in the case of Deutsche Bank Natl. Trust Co. v LeTennier, 2026 NY Slip Op 00040 (3d Dep’t Decided Jan. 8, 2026). And it is absolutely worthy of reading.
LeTennier Case Context
The context of the case was a rather mundane mortgage foreclose action with a typical fact pattern: The individual defendant “executed a note to borrow a certain sum from Nexus Financial LLC. Such note was secured by a mortgage on real property located in Delaware County. Defendant defaulted on his obligations under the note and plaintiff commenced this mortgage foreclosure action… .” Thereafter, for the next seven years, the parties battled through an onslaught of motions, appeals and other proceedings, leading ultimately to judgment of foreclose and sale against the defendant and in favor of plaintiff.
The matters that gave rise to this appeal were a number of decisions of the court below adverse to the defendant. One of those decisions denied defendant’s motion to vacate an order based upon defendant’s argument of “newly discovered evidence.” The Third Department quickly and without any hesitation disposed of that aspect of the appeal, ruling (footnote omitted):
Relating to defendant’s challenge of the March 2023 order, “[a] trial court may relieve a party from a judgment or order on the basis of newly-discovered evidence which would probably have produced a different result and which could not have been discovered in time [before disposition] or on the basis of fraud, misrepresentation, or other misconduct of an adverse party” (HSBC Bank USA, N.A. v Sage, 143 AD3d 1214, 1215 [3d Dept 2016] [internal quotation marks, ellipsis and citations omitted]). In doing so, “[a] party seeking relief from a judgment upon the ground of newly discovered evidence bears the burden of demonstrating that such proof could not have been discovered sooner through the exercise of due diligence” (Matter of Romine v New York Pub. Serv. Commn., 209 AD3d 1197, 1198 [3d Dept 2022] [internal quotation marks and citations omitted], appeal dismissed 39 NY3d 1060 [2023]). Here, defendant offered certain evidence from the US Securities and Exchange Commission (hereinafter SEC) and the agencies associated with the Secretary of State in California and New York, but failed to demonstrate why this evidence was not available or could not, with due diligence, have been discovered at the time that he opposed plaintiff’s motion for summary judgment (see Wall St. Mtge. Bankers, Ltd. v Rodgers, 148 AD3d 1088, 1089 [2d Dept 2017]; HSBC Bank USA, N.A. v Sage, 143 AD3d at 1215). Indeed, the certificate from California indicated it was issued more than a year before summary judgment was granted, and the correspondence from the SEC indicated the records were also publicly available online — which information proffered by defendant on this motion was from 2007-2008. As to his allegations of fraud and other misconduct on the part of plaintiff, such contentions are conclusory and otherwise unsupported by the record (see Carlson v Dorsey, 161 AD3d 1317, 1320 [3d Dept 2018]; Wells Fargo, N.A. v Levin, 101 AD3d 1519, 1521 [3d Dept 2012], lv dismissed 21 NY3d 887 [2013]). Nevertheless, the gravamen of defendant’s contentions distill to another challenge to plaintiff’s standing to maintain this action, which this Court has already addressed and remains the law of the case (see 189 AD3d at 2024-2025). Since defendant failed to raise any challenge concerning the August 2024 order or the September 2024 order in his initial brief, his appeals from such orders are deemed abandoned (see Matter of Shannon, 240 AD3d 1021, 1022 [3d Dept 2025]; Amici v Mazza, 234 AD3d 1170, 1172 n 2 [3d Dept 2025], lv denied 44 NY3d 902 [2025]).
The Court then went on to address what it observed was the “unconventional” aspect of the appeal: The unabashed abuse of AI.
GenAI Hallucinations
Anyone but an ostrich would be aware by now that Artificial Intelligence (AI) and generative artificial intelligence (GenAI) have not only proliferated in every aspect of life, but have exploded into the legal world with a vengeance—much of it producing an abundance of remarkable benefits. When it comes to using these new AI powers, I am reminded of what I have observed about the cause of action for fraud itself: What Do Fraud and Spiderman Have in Common? With Great Power Comes Great Responsibilities. Yes, with great AI powers, indeed come great responsibilities.
There has been a fair amount of debate as to whether the use of AI in legal research or analysis has a proper role. Some have even suggested banning such use. That is and would be, in my opinion, an entirely unnecessary and unworkable edict. As New York Supreme Court, Commercial Division, Justice Timothy S. Driscoll observes in his entertaining and enlightening seminars on the topic, AI does in fact have tremendous beneficial uses in both life and the legal profession—but nevertheless requires a heavy dose of President Ronald Reagan’s immortal cautionary adage: “Trust, but Verify.”
The decision of Justice Lisa M. Fisher in the LeTennier case is a thoughtful and instructive assessment of both the benefits and potentially menacing consequences of GenAI as used in legal advocacy.
As to acceptance of AI in legal research, Justice Fisher observed: “Although other state and federal courts have had the occasion to address the misuse of AI in legal papers, this Court has not yet had that opportunity. In doing so, we recognize that, as did the shift from digest books to online legal databases, generative artificial intelligence (hereinafter GenAI) represents a new paradigm for the legal profession, one which is not inherently improper, but rather has the potential to offer benefits to attorneys and the public — particularly in promoting access to justice, saving costs for clients and assisting courts with efficient and accurate administration of justice [citations and footnote omitted].”
Justice Fisher then recognized the mischief that can accompany this new technologic tool: “At the same time, attorneys and litigants must be aware of the dangers that GenAI presents to the legal profession. At the forefront of that peril are AI ‘hallucinations,’ which occur when an AI database generates incorrect or misleading sources of information due to a ‘variety of factors, including insufficient training data, incorrect assumptions made by the model, or biases in the data used to train the model’ … Hallucinated cases may look like a real case because they include familiar-looking reporter information, but their citations lead to cases with different names, in different courts and on different topics — or even to no case at all …. Even where GenAI provides accurate case citations, it nonetheless may misrepresent the holdings of the cited cases — often in favor of the user supplying the query [citations and footnotes omitted].”
Justice Fisher recounted the abuse of GenAI in the case before the Court in graphic terms:
[D]efendant’s opening brief cites six cases which do not exist. Plaintiff identified these fabricated cases as possibly being the product of artificial intelligence (hereinafter AI), and moved for an order seeking, among other things, sanctions against defendant and defense counsel. Defendant claimed the nonexistent cases were citation or formatting errors that he would correct in his reply brief, and then opposed the motion for sanctions with more fake cases and interpretations for existing cases that are at best strenuously attenuated, and at worst entirely inapposite. Defendant’s subsequent reply brief acknowledged that his “citation of fictitious cases is a serious error” and that they are “problematic,” but failed to offer any corrections or further explanation as previously stated. He then proceeded to include more fake cases and false legal propositions in two subsequent letters to this Court that requested judicial notice of a bankruptcy stay. In examining the propriety of defendant’s previously filed papers, more nonexistent cases were discovered in a motion that granted affirmative relief to defendant. Defense counsel reluctantly conceded during oral argument that he used AI in the preparation of his papers and, although he told the Court that he checked his papers, the filings themselves demonstrate otherwise. In total, defendant’s five filings during this appeal include no less than 23 fabricated cases, as well as many other blatant misrepresentations of fact or law from actual cases.
Sanctions Imposed
The Court found authority for imposing sanctions in the existing New York rules and cases: “In New York, courts have discretion to award costs or impose financial sanctions against an attorney or party for engaging in frivolous conduct (see 22 NYCRR 130-1.1 [a], [b]; WFE Ventures, Inc. v GBD Lake Placid, LLC, 197 AD3d 824, 836 [3d Dept 2021]).”
In applying these principles as well as the New York Rules of Professional Conduct to the facts, the Court observed (footnotes omitted):
Here, defendant submitted at least 23 fabricated legal authorities across five filings during the pendency of this appeal. He has also misrepresented the holdings of several real cases as being dispositive in his favor — when they were not. It is axiomatic that submission of fabricated legal authorities is completely without merit in law and therefore constitutes frivolous conduct (see 22 NYCRR 130-1.1 [c] [1]). It cannot be said that fabricated legal authorities constitute “existing law” so as to provide a nonfrivolous ground for extending, modifying or reversing existing law (see 22 NYCRR 130-1.1 [c] [1]; see generally Wadsworth v Walmart Inc., 348 FRD at 495; Mata v Avianca, Inc., 678 F Supp 3d at 461). Defense counsel acknowledged during oral argument that the papers were his own, and, nevertheless, “[b]y signing a paper, an attorney or party certifies that, to the best of that person’s knowledge, information and belief, formed after an inquiry reasonable under the circumstances . . . the presentation of the paper or the contentions therein are not frivolous” (22 NYCRR 130-1.1a [b] [1]). However, the excuse for fabricated legal authorities offered by defendant in his papers is incredible. Moreover, during oral argument defense counsel estimated that 90% of the citations he used were accurate, which, even if it were true, is simply unacceptable by any measure of candor to any court (see Rules of Prof Conduct [22 NYCRR 1200.0] rule 3.3 [a] [1]). Where we are most troubled is that more than half of the fake cases offered by defendant came after he was on notice of such issue, whereby his reliance on fabricated legal authorities grew more prolific as this appeal proceeded — despite it being apparent to him that such conduct lacked a legal basis (see Ader v Ader, 87 Misc 3d 1213[A], *1; see also 22 NYCRR 130-1.1 [c]; Rules of Prof Conduct [22 NYCRR 1200.0] rule 3.3 [a] [1]). Rather than taking remedial measures or expressing remorse, defense counsel essentially doubled down during oral argument on his reliance of fake legal authorities as not “germane” to the appeal.
With Great AI Powers, Come Great Responsibilities
The Court then imposed a sanction of $5,000 on defense counsel for the AI abuse (adding additional amounts for the substance of the appeal itself), and as I have observed with fraud and Spiderman, it cautioned others that with great AI powers, come great responsibilities: “To be clear, attorneys and litigants are not prohibited from using GenAI to assist with the preparation of court submissions. The issue arises when attorneys and staff are not sufficiently trained on the dangers of such technology, and instead [*7] erroneously rely on it without human oversight. As with the work from a paralegal, intern or another attorney, the use of GenAI in no way abrogates an attorney’s or litigant’s obligation to fact check and cite check every document filed with a court. To do otherwise may be sanctionable, depending on the facts and particular circumstances of each case.”