Joseph R. Saveri

San Francisco
Known for successfully bringing class actions for plaintiffs ranging from electronics manufacturers to cheerleaders' families, Joseph Saveri has long believed in pushing the law into new areas. "There's some value in advancing the law in ways where there haven't been cases or where the law isn't particularly well developed," he said.
In the last two years, he and the eponymous firm he founded in 2012 have been pushing the law to find value in the works of creators that artificial intelligence companies are using to train their large language models. Beginning in November 2022, he has brought copyright and other class actions against seven generative AI companies on behalf of writers, visual artists and software developers.
"All the cases share a basic premise, which is that these companies ... in order to produce those [AI] products, used or took training data that was owned by others," Saveri said. "They acquired them sometimes through means that aren't really clear and are probably not legal."
"Our claim at least is that the copying of that material, which was protected by our client's copyrights, is illegal, violates the Copyright Act and violates perhaps other statutes."
"We've had a lot of copycats following our same theory, including companies like the New York Times," he added.
Saveri filed his first AI class action on behalf of computer programmers against a product that produces computer code. That case does not raise copyright claims, but "that means that the defendants don't have a fair use defense," he said. Doe 1 v. GitHub Inc., 4:22-cv-06823 (N.D. Cal., filed Nov. 3, 2022).
The judge recently dismissed the plaintiffs' claims under the Digital Millenium Copyright Act, but Saveri is considering appealing.
"Our theories when we started are broad and untested. ... It's fair to say that we've had some successes, and all of our cases are proceeding into discovery," he said. "But we've also lost some of the motions to dismiss. And the courts have found ... some of our theories to be good and others not so good."
The case with the largest potential impact is against the creator of ChatGPT, OpenAI.Tremblay v. OpenAI Inc., 3:23-cv-03223 (N.D. Cal., filed June 28, 2023).
District Judge Araceli Martínez-Olguin has kept the case moving and in February dismissed a number of claims. But now, Saveri said, his team is deep into discovery. "We find that the evidence we've seen so far really provides a lot of support for ... the detailed allegations of the complaint," he said. "So we think the case is shaping up."
His newest class action is against Nvidia, famous for its AI chips, over the company's own generative product. Nazemian v. NVIDIA Corp., 4:24-cv-01454 (N.D. Cal., filed May 8, 2024).
But there will be more, he promised. "With the passage of time, and ... [as] we keep pushing this rock up the hill, I think people are coming to the conclusion that we're right."