New York, March 26, 2025 — In a highly unusual courtroom moment that quickly went viral, a New York appellate court judge abruptly stopped a hearing after discovering that the “lawyer” arguing on video was not a real person, but an AI-generated avatar.
The incident occurred during a session of the New York State Supreme Court Appellate Division, First Department. Jerome Dewald, a 74-year-old self-represented plaintiff (pro se) in an employment dispute, had submitted a pre-recorded video argument as permitted by the court. However, what played on screen was not Dewald himself, but a polished, professional-looking young man in a home-office setting, speaking eloquently on his behalf.
Almost immediately after the video began, Justice Sallie Manzanet-Daniels interrupted.
“Hold on… Is that counsel for the case?” she asked.
Dewald, who was present in the courtroom, admitted: “I generated that. That’s not a real person.”
The judge expressed clear displeasure, noting that Dewald had not disclosed the use of artificial intelligence in his application for video submission. The video was promptly shut off. Court observers described the moment as one of visible judicial frustration, with the panel emphasizing that the courtroom was not the place to test or promote AI legal tools.
Dewald later acknowledged that the court “chewed me up pretty good.” He subsequently submitted a letter of apology to the justices. In interviews following the hearing, he explained that he used the AI avatar due to personal speech difficulties and a desire to present a clearer, more professional argument. He is also a tech entrepreneur who has been developing and promoting AI tools for legal applications.
Why the Strong Reaction?
Legal experts point to several critical issues raised by the case:
- Lack of Transparency: While Dewald had permission to submit a video, he failed to inform the court that the speaker was AI-generated. Courts demand full disclosure, especially regarding representation.
- Ethical and Professional Standards: Only licensed attorneys (or self-represented litigants within strict limits) may argue cases. An AI avatar is not a licensed lawyer, is not an officer of the court, and cannot be held accountable for ethical violations, contempt of court, or misleading statements in the same way a human attorney can.
- Broader Implications for AI in Law: This was not the first clash between AI and courtroom rules. Previous cases have seen lawyers sanctioned for submitting briefs with fake citations generated by ChatGPT. The incident underscores growing concerns about the appropriate boundaries for AI use in the justice system.
Dr. Adam Wandt, an expert on technology and law, noted that while AI tools are increasingly used for legal research, drafting, and case analysis, judges are highly unlikely to allow AI avatars to serve as virtual counsel in the foreseeable future.
Aftermath and Significance
Dewald continued arguing his case in person after the video was stopped. The event, however, captured widespread attention online, with clips of the hearing circulating widely on social media. It has fueled ongoing debates about the rapid integration of artificial intelligence into traditionally human domains like legal advocacy.
As courts worldwide grapple with the rise of generative AI, this New York hearing serves as a stark reminder: while technology may evolve quickly, the foundational principles of transparency, accountability, and human responsibility in the courtroom remain firmly in place.
About The Author
