Jerome Dewald sat upright in front of five New York State appellate judges. A 74-year-old entrepreneur without legal training, he was there to contest a ruling in his employment case. But the real twist wasn’t in the lawsuit. It came when the video he submitted began playing in court.
A young, clean-cut man appeared on the courtroom screen. “May it please the court,” the digital figure began. “I come here today a humble pro se before a panel of five distinguished justices.”
It took just seconds for Justice Sallie Manzanet-Daniels to interrupt. “Ok, hold on. Is that counsel for the case?”
“No,” Dewald replied from his seat. “I generated that. That’s not a real person.”
“I don’t appreciate being misled”
The courtroom fell silent. The judge, visibly irked, responded sharply.
“It would have been nice to know that when you made your application. You did not tell me that, sir,” said Justice Manzanet-Daniels. She then raised her voice across the courtroom: “Shut that off. I don’t appreciate being misled.”
The moment, captured on court cameras, has since stirred debate about the growing influence—and misuse—of AI in legal spaces.
Why an AI Avatar?
In an interview with The Associated Press, Dewald explained his reasoning. Representing himself in court had proven difficult. His voice often faltered under pressure. So, he applied to submit a pre-recorded video—and used AI to deliver it more clearly.
“I generated that,” Dewald admitted again. “That is not a real person.” He had used software developed by a San Francisco tech firm to build the avatar. Although he had initially intended to create a version that resembled him, technical issues got in the way.
“The court was really upset about it,” he said. “They chewed me up pretty good.”
In a letter to the court, Dewald later wrote, “My intent was never to deceive but rather to present my arguments in the most efficient manner possible. However, I recognise that proper disclosure and transparency must always take precedence.”
He added that the avatar was meant to compensate for his nervous speech: “I stumbled a lot in earlier hearings. I thought this might help.”
Cautionary tale in the courtroom
After the video was stopped, Dewald continued his argument the old-fashioned way. He spoke slowly, pausing often, and read from his phone.
His embarrassment, however, isn’t unique. Several licensed attorneys have also faced trouble for mishandling artificial intelligence in court.
In June 2023, two New York lawyers and their firm were fined $5,000 each after submitting a legal brief filled with fabricated cases. They had relied on an AI chatbot, unaware that it could invent citations. The firm claimed it was a “good faith mistake”.
Later that year, former Trump lawyer Michael Cohen landed in similar trouble. He passed on legal citations to his attorney—only to later discover they had been hallucinated by Google Bard. “I didn’t know it could make stuff up,” Cohen pleaded in court.
The bigger picture: AI's legal footprint grows
Dewald’s case underscores a larger shift. Courts are already experimenting with AI in controlled ways. Just last month, the Arizona Supreme Court launched two AI avatars named “Daniel” and “Victoria” to summarise decisions on its public website. These, however, come with full transparency.
“From my perspective, it was inevitable,” said Daniel Shin, assistant director of research at the Centre for Legal and Court Technology at William & Mary Law School. “They can still hallucinate—produce very compelling looking information that is either fake or nonsensical. That risk has to be addressed.”
Shin noted that while licensed lawyers would likely avoid such tactics due to the risk of disbarment, individuals without legal counsel aren’t given much guidance about the risks of AI.
“There are no real instructions about this kind of thing,” he said.
Dewald said he had recently watched an American Bar Association webinar about AI in the legal world. He saw himself as simply trying to keep up. His case, still pending as of Thursday, may yet set precedent in how courts treat AI submissions—especially from those without legal representation.
But for now, the message from the bench is clear: the courtroom may be slow to adapt to avatars, even if technology continues to race ahead.
A young, clean-cut man appeared on the courtroom screen. “May it please the court,” the digital figure began. “I come here today a humble pro se before a panel of five distinguished justices.”
It took just seconds for Justice Sallie Manzanet-Daniels to interrupt. “Ok, hold on. Is that counsel for the case?”
“No,” Dewald replied from his seat. “I generated that. That’s not a real person.”
“I don’t appreciate being misled”
The courtroom fell silent. The judge, visibly irked, responded sharply.
“It would have been nice to know that when you made your application. You did not tell me that, sir,” said Justice Manzanet-Daniels. She then raised her voice across the courtroom: “Shut that off. I don’t appreciate being misled.”
The moment, captured on court cameras, has since stirred debate about the growing influence—and misuse—of AI in legal spaces.
Why an AI Avatar?
In an interview with The Associated Press, Dewald explained his reasoning. Representing himself in court had proven difficult. His voice often faltered under pressure. So, he applied to submit a pre-recorded video—and used AI to deliver it more clearly.
“I generated that,” Dewald admitted again. “That is not a real person.” He had used software developed by a San Francisco tech firm to build the avatar. Although he had initially intended to create a version that resembled him, technical issues got in the way.
“The court was really upset about it,” he said. “They chewed me up pretty good.”
In a letter to the court, Dewald later wrote, “My intent was never to deceive but rather to present my arguments in the most efficient manner possible. However, I recognise that proper disclosure and transparency must always take precedence.”
He added that the avatar was meant to compensate for his nervous speech: “I stumbled a lot in earlier hearings. I thought this might help.”
Cautionary tale in the courtroom
After the video was stopped, Dewald continued his argument the old-fashioned way. He spoke slowly, pausing often, and read from his phone.
His embarrassment, however, isn’t unique. Several licensed attorneys have also faced trouble for mishandling artificial intelligence in court.
In June 2023, two New York lawyers and their firm were fined $5,000 each after submitting a legal brief filled with fabricated cases. They had relied on an AI chatbot, unaware that it could invent citations. The firm claimed it was a “good faith mistake”.
Later that year, former Trump lawyer Michael Cohen landed in similar trouble. He passed on legal citations to his attorney—only to later discover they had been hallucinated by Google Bard. “I didn’t know it could make stuff up,” Cohen pleaded in court.
The bigger picture: AI's legal footprint grows
Dewald’s case underscores a larger shift. Courts are already experimenting with AI in controlled ways. Just last month, the Arizona Supreme Court launched two AI avatars named “Daniel” and “Victoria” to summarise decisions on its public website. These, however, come with full transparency.
“From my perspective, it was inevitable,” said Daniel Shin, assistant director of research at the Centre for Legal and Court Technology at William & Mary Law School. “They can still hallucinate—produce very compelling looking information that is either fake or nonsensical. That risk has to be addressed.”
Shin noted that while licensed lawyers would likely avoid such tactics due to the risk of disbarment, individuals without legal counsel aren’t given much guidance about the risks of AI.
“There are no real instructions about this kind of thing,” he said.
Dewald said he had recently watched an American Bar Association webinar about AI in the legal world. He saw himself as simply trying to keep up. His case, still pending as of Thursday, may yet set precedent in how courts treat AI submissions—especially from those without legal representation.
But for now, the message from the bench is clear: the courtroom may be slow to adapt to avatars, even if technology continues to race ahead.
You may also like
Pep Guardiola addressed Kevin De Bruyne 'rift' claim after Man City exit confirmed
Liz Hurley shares sweet snaps of lookalike son and pays tribute to 'sunbeam' on his birthday
All India Muslim Women Personal Law Board backs Waqf Amendment Bill
Saturday Kitchen's Matt Tebbutt says 'if it wasn't live we'd cut that out' over show blunder
Defence Minister Rajnath Singh inaugurates various infra projects at Karwar naval base
