Matthew Flanagan

The latest instance occurred in Landberg v. City of New York, 2026 NY Slip Op 03935, 2026 LX 36495 (2d Dept., June 23, 2026), which was a fairly simple trip and fall case. The plaintiff tripped and fell on a loose brick on a sidewalk and brought an action against the City of New York and the abutting landlord, both of which were awarded summary judgment. The Appellate Division, Second Department, affirmed the grant of summary judgment in a relatively short decision issued in early June. 

The court spent far more time later in the month addressing the contents of the brief submitted on behalf of the plaintiff and the conduct of plaintiff’s attorney. In a lengthy decision, the court dissected the plaintiff’s brief, finding that it cited and/or quoted from three cases which did not exist, and then misrepresented the holdings of cases which actually did exist.  One of the fictious quotes purported to be from the court of appeals, but was “directly contrary to what the court of appeals has actually said.” After being confronted with the non-existent cases at the oral argument, the plaintiff’s attorney was offered an opportunity to take a 15-minute recess to come up with an explanation of where the nonexistent cases came from, but he declined the opportunity and said that he had “found them either on Westlaw or Lexis or in a book, a previous case, or another brief.” The court concluded that the attorney was not being candid, and that he should have said (as he later admitted) “that the cases were hallucinated by artificial intelligence.”

The court imposed a monetary sanction of $8,000 on the attorney, but the court did not stop there. Although the handling attorney’s actions were contrary to the policies of the law firm he worked for, the firm was sanctioned, too (for $2,500). The attorneys for the defendants, while not sanctioned, were chastised for not alerting the court to the AI hallucinations in plaintiff’s brief.  The court wrote: “Although the attorneys for the owner and the City, who are officers of the court, submitted respondents’ briefs, neither of them alerted this Court about the fabricated citations, fabricated quotations, misstatements of law, and misrepresentations regarding the holdings of real cases contained in the plaintiff’s brief.”  

Not surprisingly, the sanction exceeded the sanction which the same court had assessed a few weeks earlier against a pro se litigant. In Matter of Julien v. Arthur, 2026 NY Slip Op 03308, 2026 LX 287806  (2d Dep.t, May 27, 2026), the Second Department sanctioned a pro se litigant in the amount of $250 for his reliance on generative AI. The pro se litigant cited only one nonexistent case, but the court ruled that that was sufficient to warrant the imposition of a sanction. The Julien court noted that the pro se litigant (a father in a custody dispute) acknowledged responsibility for his conduct and that it only occurred once, which undoubtedly factored into the lower monetary sanction. The following month, in distinguishing the much lower sanction awarded in Julien, the Landberg court said: “This case is obviously much more egregious in that the person who filed the subject brief is a trained attorney with access to legal research materials, including Westlaw, and who is bound by the Rules of Professional Conduct.”

Attorneys practicing in the Second Department can expect a sanction greater than $250 for even a single AI hallucination or fabricated citation in a brief. The court made it clear in Landberg that attorneys are on notice of the unreliability of generative AI: “stories about the unreliability of GenAI in performing legal work, leading to fictitious citations, misrepresentations of the law, and sanctions, have been widespread in the media for years.”  The court also made it clear that attorneys are expected to alert the court to the AI hallucinations and fabricated cases and quotations in their adversary’s briefs. 

The foregoing cases coincide with the New York State Unified Court System’s adoption of a new rule, Part 161, regulating the use of AI. The new rule provides that attorneys cannot be prohibited from using artificial intelligence tools in preparing papers submitted to a court, but the tools must be used “in accordance with the duties and responsibilities that apply to individuals who submit papers to a court.”

The rule goes on to suggest that individual judges should adopt a Part Rule which provides: “Every attorney or party who uses an artificial intelligence (AI) tool in preparing any paper submitted to this court is expected to understand that tool's capabilities and limitations. Attorneys and parties need to be aware that AI tools, among other risks and limitations, can generate fabricated information or fictitious citations to authority (commonly known as hallucinations)… Accordingly, any attorney or party who uses an artificial intelligence tool, as defined in 22 NYCRR 161.2(a), in preparing any paper, as defined in 22 NYCRR 161.2(b), filed in or submitted to this court or served on another party in a case before this court is required to carefully review the paper and independently ensure that it contains no fabricated or fictitious cases, statutes, or other material.”

Attorneys are on notice of the pitfalls of blind reliance on generative AI and can expect to be sanctioned if they mis-cite or misquote cases in submissions which are prepared with the assistance of AI tools.