Comedian and author Sarah Silverman, along with authors Christopher Golden and Richard Kadrey — are suing U.S. District Court for copyright infringement The double accusation sues OpenAI and Meta respectively.
The lawsuit alleges that, among other things, OpenAI’s ChatGPT and Meta’s LLaMA were trained on illegally obtained datasets containing their work, which they say were derived from Bibliotik, Library Genesis, and others ” Z-Library, etc., and notes that the books are “available in bulk through the torrent system”.
Both Golden and Kadrey declined to comment on the lawsuit, while Silverman’s team did not respond by press time.
In the OpenAI lawsuit, the three presented evidence shows that, when prompted, ChatGPT summarizes their book, infringing their copyright. Silverman’s Bedwetter was the first book ChatGPT showed in the exhibit, while Golden’s book Ararat was also used as For example, Kadrey’s book Sandman Slim. The claim states that the chatbot never bothered to “copy any copyright management information that the plaintiffs included in its published works.”
As for the separate lawsuit against Meta, it claims that it could be used to train its Access the author’s book on the dataset for the LLaMA model, four open-source AI models the company launched in February.
The complaint steps through why plaintiffs believe the data sets have an illicit origin—in a meta-paper detailing LLaMA, the company states that its The sources of the training datasets, one of which is called ThePile, was assembled by a company called EleutherAI. The complaint states that a paper by EleutherAI described ThePile as being cobbled together from “copies of the contents of Bibliotik’s private tracker.” The suit says Bibliotik and the other “shadow libraries” listed are “blatantly illegal.”
In both claims, the authors said they “did not consent to the use of their copyrighted book as training material for the company’s artificial intelligence models.” Their lawsuits each contain six different counts of copyright infringement, negligence, unjust enrichment and unfair competition. The author is seeking statutory damages, return of profits, etc.
Attorneys Joseph Saveri and Matthew Butterick, representing the three authors, wrote on their LL.M. litigation website that they stemmed from “concerns about [ChatGPT] generating similar copyrighted textual material (including digital writers, authors, and publishers of uncanny ability to write texts in thousands of books.”
Saveri has also begun litigation against artificial intelligence companies on behalf of programmers and artists. Getty Images has also filed an AI lawsuit alleging that Stability AI, which created the AI image generation tool Stable Diffusion, trained its models on “millions of copyrighted images.” Saveri and Butterick also represented authors Mona Awad and Paul Tremblay in a similar case involving the company’s chatbot.
Lawsuits like this are not only a headache for OpenAI and other AI companies, but also for OpenAI. They are pushing the limits of copyright. As we said on The Vergecast, whenever someone gets Nilay involved in copyright law, we’re going to see lawsuits around this stuff in the next few years.
We reached out to Meta, OpenAI, and the law firm of Joseph Saveri for comment, but they had not responded by press time.
The following is the lawsuit: