Jerome Dewald just tried to argue his case in a New York courtroom—via a generative AI lawyer avatar. Yes, really.
The 74-year-old engineer and AI entrepreneur, who represents himself in court, got prior approval to show a video during his hearing. What he didn’t mention? That the well-groomed man in a V-neck sweater delivering the opening lines—“I come here today a humble pro se before a panel of five distinguished justices…”—wasn’t real.
The avatar, named James (but let’s call him Jim, because of course), was generated using AI tool Tavus. Within seconds of the video starting, Associate Justice Sallie Manzanet-Daniels interrupted with the legal equivalent of: “Wait… who the hell is that?”
“That? I generated that,” Dewald told her.
“That is not a real person.”
See also: No, Scientists Didn’t Resurrect the Dire Wolf — But What They Did Do Is Even Weirder
When AI Lawyer Meets the Real World
The courtroom was not amused. Dewald was immediately chastised for using the court as a product demo for his startup, Pro Se Pro, which promises AI-based legal aid for people representing themselves. Dewald insists it wasn’t a business stunt—just a workaround due to long-standing throat issues following cancer.
Still, the judge didn’t buy it, pointing out that he’d verbally spoken to staff in recent days just fine. The trial was for an employment dispute with insurance firm MassMutual, but by the time Jim was booted offscreen, the legal argument had morphed into a strange meditation on technology, legitimacy, and who gets to “speak” in court.
In a scene better suited to Black Mirror than a New York appeals court, 74-year-old retiree Jerome Dewald tried to argue his employment case by playing a video—featuring not himself, but a sleek AI-generated avatar delivering legal arguments on his behalf. Dewald, representing… pic.twitter.com/4gT92TUru3
— litigation_god (@GodLitigation) April 10, 2025
“Jim” Was Not the Plan—But He Was Hot
Dewald says he meant to create a digital replica of himself, but his AI budget only allowed for three renders per month, and they all failed. So he used a stock avatar from Tavus:
“That big, beautiful hunk of a guy they call Jim.”
Not a joke. That’s a direct quote.
Jim got about 20 seconds of screen time before being cut off—presumably before he could cite any hallucinated legal precedents.
See also: Doctors Claim They Can Remove Microplastics From Your Blood—But Does It Work?
When AI Enters Courtrooms, Who Gets to Speak?
There are two big questions here:
One: Are we overreacting to a guy using ChatGPT in court, or underreacting to what this means for the future of legal accountability?
Two: If this becomes normalized, what human professions are next to be turned into prompt templates?
This is bigger than one courtroom stunt. Dewald isn’t some rogue hacker in a basement—he’s part of a growing wave of tech optimists who believe AI can democratize access to justice. And in a country where public defenders are overworked and representation is a luxury, they’re not entirely wrong.
But replacing lived legal expertise with avatars—ones that might misstate facts, fabricate citations, or (God help us) base arguments on vibes—doesn’t exactly scream “equal access.” It screams Studio Ghibli fever dream meets Black Mirror court drama.
While we are flattered that people think so highly of our replicas, please do not use them in court (yet) 🙃https://t.co/UMjnNLSq0D
— behr (@axbehr) April 4, 2025
What’s Next—AI Therapists? AI Lovers? AI Babysitters?
We’ve already seen art, music, journalism, and animation swept up in generative AI. (Ghibli-style animation filters, anyone?) Courtrooms are just the next frontier.
And if avatars like Jim become the norm, what happens to real human voices—especially those already pushed to the periphery?
Dewald, to his credit, filed an apology. He insists the tech empowers marginalized people to be heard. But even he admits that “full and open disclosure” about using AI still triggers bias. People assume if AI is involved, it’s a lie.
And sometimes… it is.
