Federal judges meeting in Boston this week heard a blunt warning from the tech industry: AI-generated images and video are becoming so convincing that there may be no reliable way left to prove they’re fake.
The warning came at a meeting of the U.S. Judicial Conference’s Advisory Committee on Evidence Rules, which has spent three years since 2023 studying whether courts need new rules for handling evidence that may have been altered or fabricated with artificial intelligence, according to Reuters.
Representatives from OpenAI and Meta told the committee on Thursday that detection technology is losing ground to generation technology, not catching up to it.
“AI images are only getting better, and I do not think that there will be a scientific way going forward to detect them,” said Shira Anderson, OpenAI’s policy counsel.
Devin Norder, Meta’s director of product management, said the newest AI models produce content that people can no longer distinguish from authentic footage or photos. “Media generated from frontier models, at least in the last few generations, are indistinguishable by humans from authentic or AI-generated media,” Norder told the panel. He noted that the barrier to making convincing fakes has essentially disappeared: “And the cost and the time it takes to create this media really are quite trivial at this point.”
Watermarking Has Limits
Both OpenAI and Meta have built their own tools to help verify whether visual content is authentic, and regulators have begun requiring similar safeguards. California and the European Union have each passed laws mandating watermarking systems meant to flag AI-generated material. But Anderson said no single approach will solve the problem, in part because different companies’ verification tools don’t communicate with one another. “My guess is that this absolutely will continue to get better. I do not think it’ll ever be bulletproof,” she said.
Few Cases So Far
Despite the alarm, the committee has found scant evidence that deepfakes are actually turning up in courtrooms yet. A March survey by the Federal Judiciary Center, the judiciary’s research arm, polled 931 judges and found only 15 had handled a case where someone challenged evidence as a deepfake.
US District Judge Jesse Furman of Manhattan, who chairs the committee, pointed to that disconnect between the dire predictions and the current reality. “There’s a lot of academic commentary and other commentary predicting that courts will be dealing with deepfakes, but we haven’t actually seen a whole lot of that yet,” Furman said.
That gap has divided the committee over whether a new, AI-specific evidence rule is even necessary. Manhattan US District Judge Valerie Caproni argued existing legal tools are sufficient. “We’ve dealt with bad actors presenting fake evidence since Cain and Abel. That’s a longstanding problem,” Caproni said.
A Growing Cost for Ordinary People
Outside the courtroom, AI-generated fakes are already being used against the public in fraud schemes. Author Wynton Hall, whose book Code Red: The Left, the Right, China, and the Race to Control AI made the New York Times best-seller list, described the tactic in a Newsmax appearance.
“They can now mine your social media and then they build a detailed dossier on your life, and then they use the AI to mimic a loved one so convincingly that your instincts will tell you that it’s real,” Hall said. “These tools are now becoming democratized so that they’re cheap, they’re easy to use, and that means they’re accessible at a mass scale.”
The FBI recorded more than 22,000 complaints tied to AI-related scams in 2025, with reported losses of nearly $893 million.
