Lawyers across the country are getting busted for using AI to write their legal briefs — and their excuses are even more creative than the fake cases they’ve allegedly been citing.
From blaming hackers to claiming that toggling between windows is just too hard, attorneys are desperately trying to dodge sanctions for a tidal wave of AI-generated nonsense clogging up court dockets.
But judges are tired of hearing it and a group of “legal vigilantes” is making sure none of these blunders go unnoticed.
A network of lawyers has been tracking down every instance of AI misuse they can find, compiling them in a public database that has swelled to over 500 cases.
The database maintained by France-based lawyer and researcher Damien Charlotin exposes fake case citations, bogus quotes and the attorneys responsible — hoping to shame the profession into cleaning up its act.
The number of cases keeps growing, Charlotin told The Post on Wednesday.
“[T]his has accelerated exactly at the moment I started cataloguing these cases, from maybe a handful a month to two or three a day,” he said in an email.
“I think this will continue to grow for a time,” Charlotin added.
He said some examples are just mistakes, and “hopefully awareness will reduce them, but that’s not a given.”
In other instances, AI is misused by “reckless, sloppy attorneys or vexatious litigants,” the researcher wrote.
“I am afraid there is little stopping them,” he added.
Amir Mostafavi, a Los Angeles-area attorney, was recently slapped with a $10,000 fine after filing an appeal in which 21 of 23 case quotes were completely made up by ChatGPT.
His excuse? He said he wrote the appeal himself and just asked ChatGPT to “try and improve it,” not knowing it would add fake citations.
“In the meantime we’re going to have some victims, we’re going to have some damages, we’re going to have some wreckages,” Mostafavi told CalMatters.
“I hope this example will help others not fall into the hole. I’m paying the price.”
Ars Technica reported that Innocent Chinweze, a New York City-based lawyer, was recently caught filing a brief riddled with fake cases. He said he’d used Microsoft Copilot for the job.
Then, in a bizarre pivot, he claimed his computer had been hacked and that malware was the real culprit.
The judge, Kimon C. Thermos, called the excuse an “incredible and unsupported statement.”
After a lunch break, Chinweze “dramatically” changed his story again — this time by claiming that he didn’t know AI could make things up.
Chinweze was fined $1,000 and referred to a grievance committee for conduct that “seriously implicated his honesty, trustworthiness, and fitness to practice law.”
Another lawyer, Alabama attorney James A. Johnson, blamed his “embarrassing mistake” on the sheer difficulty of using a laptop, according to Ars Technica.
He said he was at a hospital with a sick family member and under “time pressure and difficult personal circumstance.”
Instead of using a bar-provided legal research tool, he opted for a Microsoft Word plug-in called Ghostwriter Legal because, he claimed, it was “tedious to toggle back and forth between programs on [his] laptop with the touchpad.”
Judge Terry F. Moorer was unimpressed, noting that Ghostwriter clearly stated it used ChatGPT.
Johnson’s client was even less impressed, firing him on the spot. The judge hit the attorney with a $5,000 fine, ruling his laziness was “tantamount to bad faith.”
Such cases are “damaging the reputation of the bar,” tephen Gillers, an ethics professor at New York University School of Law, told the New York Times.
“Lawyers everywhere should be ashamed of what members of their profession are doing,” he added.
Still, the excuses for AI mistakes keep coming. One lawyer blamed his client for helping draft a problematic filing. Another claimed she had “login issues with her Westlaw subscription.”
A Georgia lawyer insisted she’d “accidentally filed a rough draft.”
But the penalties are getting steeper. Florida lawyer James Martin Paul was reportedly hit with a staggering $85,000 sanction for “repeated, abusive, bad-faith conduct that cannot be recognized as legitimate legal practice and must be deterred.”
When he argued the fine was too high, the court shot back that caving to his arguments “would only benefit serial hallucinators.”
Illinois attorney William T. Panichi has been sanctioned at least three times, Ars Technica found.
After the first, he promised the court, “I’m not going to do it again,” just before getting hit with two more rounds of sanctions a month later.
Judges are losing their patience.
“At this point, to be blunt, any lawyer unaware that using generative AI platforms to do legal research is playing with fire is living in a cloud,” wrote US Bankruptcy Judge Michael B. Slade.
Another judge, Nancy Miller, blasted a lawyer who argued it only takes “7.6 seconds” to check a citation. Miller noted that the lawyer herself had failed to take those “precious seconds” to check her own work.
As one Texas judge put it, “At one of the busiest court dockets in the nation, there are scant resources to spare ferreting out erroneous AI citations.”
The Post has sought comment from Mostafavi, Chinweze, Johnson, Paul and Panichi.
Credit: Source link
