10
Loading video...
By Shannon Power
Pop Culture & Entertainment Reporter
0
Share
Newsweek is a Trust Project member
See more of our trusted coverage when you search.Prefer Newsweek on Googleto see more of our trusted coverage when you search.
Sarah Silverman has suffered a setback in her lawsuit against Meta, the company that owns Facebook, over Artificial Intelligence (AI).
Silverman sued the company over what she claimed was unauthorized use of her copyrighted books to train its generative AI mode.
But U.S. District Judge Vince Chhabria threw out much of Silverman's lawsuit in a ruling on Monday, describing some of the claims as "nonsensical."
Read More
Selena Gomez Is in Hot Water4 min read
Is Amy Schumer Really the Most Successful Female Comedian of All Time?6 min read
Trump-Biden Polling Traumatizes Late Night Hosts4 min read
...
"There is no way to understand the LLaMA models themselves as a recasting or adaptation of any of the plaintiffs' books," Chhabria wrote in his ruling, as per The Hollywood Reporter. LLaMA refers to the Large Language Model Meta AI.
The judge also dismissed the argument that all results from Meta's AI tool would constitute copyright infringement as Silverman did not provide the evidence that "could be understood as recasting, transforming, or adapting the plaintiffs' books."
Newsweek contacted Silverman's representatives and Meta by email for comment.
Silverman's lawsuit also argued that using her IP was resulting in infringing derivatives because those companies would benefit from every search by third-party users. It also accused the AI companies of violating competition laws.
"To prevail on a theory that LLaMA's outputs constitute derivative infringement, the plaintiffs would indeed need to allege and ultimately prove that the outputs 'incorporate in some form a portion of' the plaintiffs' books," Chhabria wrote in his ruling.
It is not the first lawsuit of its kind to argue that intellectual property such as books, art and images used to train AI bots constituted copyright infringement. Silverman joined that class action against Meta in July, but U.S. District Judge William Orrick also dismissed many arguments in that lawsuit.
He labeled the arguments as "defective in numerous respects," because the plaintiffs could not provide proof of material produced by AI.
That suit also sued StabilityAI, DeviantArt and Midjourney.
Meta did not dismiss the claim that using books to train its AI bots is comparable to copyright infringement.
Many AI companies also argue they do not need to get licenses because they are allowed to access the content under the protection of fair use.
As a result of the two recent rulings, plaintiffs will likely have to provide evidence of copyright infringement created by AI machines that is identical to the original material.
To successfully sue under copyright law, a test of substantial similarity is used to measure potential infringement, which some experts say is unlikely to occur with AI.
View 0 Comments
Submit Correction
View Editorial & AI Guidelines