The Second Department has spoken, and the message is unmistakable: AI is a tool—not an excuse.
The Court imposed an $8,000 sanction against the attorney who filed an appellate brief containing fabricated case citations, fictional quotations attributed to the Court of Appeals, and misrepresentations of what actual cases held. The attorney's law firm was also sanctioned $2,500.
The opinion is worth reading, not just for its discussion of AI hallucinations, but for what ultimately drove the Court's decision. Landberg v City of New York - 2026 NY Slip Op 03935.
When questioned at oral argument about the bogus authorities, the attorney told the Court they came from Lexis, Westlaw, a book, or another brief. Later, in response to the Court's Order to Show Cause, he admitted they were likely generated by one of the free AI tools and acknowledged that he was "scared" to say so during oral argument.
The Court didn't mince words. It found that he was not candid with the Court and described his conduct as demonstrating "professional immaturity, arrogance, and profound lack of respect for the judicial system."
Those are words no lawyer ever wants to see attached to his or her name.
One aspect of the opinion that caught my attention involves the respondents' attorneys.
At oral argument, the panel questioned why neither respondent had alerted the Court that the appellant's brief contained fabricated cases and quotations. The Court plainly expressed its disappointment. Yet, in the end, the respondents escaped sanctions. Because neither defendant sought costs or submitted billing records, the Court declined to award them.
That leaves an interesting question for practitioners.
When opposing counsel files a brief containing hallucinated authorities, is it enough to simply point out the errors in your responsive brief? Or does your duty as an officer of the court require you to affirmatively notify the Court before argument?
The Second Department stopped short of answering that question directly. But the discussion at oral argument—and the Court's treatment of the issue in its opinion—suggests that judges expect more than silence.
The broader lesson is clear.
The Court made a point of saying that lawyers are not prohibited from using generative AI. Quite the opposite. AI can be an extraordinarily useful tool.
What lawyers cannot do is delegate their professional judgment to it.
Every citation must be checked.
Every quotation must be verified.
Every proposition of law must be confirmed against an actual source.
Even if you are acting as local counsel for someone, you are exposed if you argue a motion or an appeal you did not write and may have to answer for an AI hallucinated case.
Those obligations existed long before generative AI. They haven't changed. The technology has.
This decision is another reminder that courts are not sanctioning lawyers for using AI. They are sanctioning lawyers for failing to practice law.
The Second Department’s holding tracks with the New York State Unified Court System’s rule on the use of AI in court submissions: AI is permissible, but the attorney is charged with making sure that their papers are accurate.
As AI becomes a routine part of legal practice, this opinion should become required reading for every litigator.
As artificial intelligence continues to reshape the legal profession, attorneys, law firms, and legal departments should ensure they have appropriate policies, training, and safeguards in place to use these tools responsibly. If you have questions about the ethical, professional, or litigation-related risks associated with AI, or would like assistance developing best practices for your organization, please contact our team. We are closely monitoring developments in this rapidly evolving area and are available to help navigate the opportunities, and pitfalls, of AI in the practice of law.
Tags:
