Comedian Sarah Silverman participated in a class action lawsuit against OpenAI and Meta, accusing each company of “copying and incorporating” her protected work to train artificial intelligence programs, alleging copyright infringement, according to court documents. sued both companies.
of lawsuitamong them she participated writers Christopher Golden and Richard Cudleywas filed Friday in the San Francisco Division of the United States District Court for the Northern District of California. Each lawsuit alleges that the company in question used illegal online scraping to make copies of the author’s works, including Silverman’s memoir “The Bedwetter,” without permission.shadow library” contains texts from thousands of books.
The lawsuit against Meta cites the company’s allegations. own research paper About LLaMA, a large-scale language model used to train chatbots. In a paper published in February, scientists included the following documents: The Pile in the training dataset. Some of the text comes from the Shadow Library, according to the lawsuit.
“Their copyrighted material was copied and ingested as part of their training,” the lawsuit alleges. “Many of the plaintiffs’ books are included in the data sets Meta has approved for use.”
Neither OpenAI nor Meta responded to a request for comment on the lawsuit on Monday. Plaintiffs seek damages and injunctive relief, which may include changes to the LLaMA and ChatGPT programs.
Not much is known about the source of the training dataset for OpenAI’s ChatGPT program. But the lawsuit states that ChatGPT’s ability to generate summaries of plaintiffs’ works “is possible only if ChatGPT is trained on plaintiffs’ copyrighted works.”
Exhibits include the text generated when asked to summarize Silberman’s memoir “The Bedwetter”.
“One of the key topics of the first part of the memoir is Silberman’s struggle with nocturnal enuresis, or nocturnal enuresis, which continued into his teenage years,” the program writes. “This issue caused her significant distress and embarrassment, but at the same time stimulated her resilience and her ability to cope with adversity.”
Attorneys representing the three authors, Joseph Saberi and Matthew Buterick, represent other creators in a separate lawsuit against Copilot, an artificial intelligence-powered coding assistant and Stability AI-generated image generator on GitHub. people are also working.
and website Publishing the lawsuit against the AI company, the attorneys said, “Much of the material in the training datasets used by OpenAI and Meta comes from copyrighted works, including books written by the plaintiffs, which are open to OpenAI. and Meta copied without consent or credit.” without compensation. “
This double action could define the boundaries of how artificial intelligence learns and what role copyright law plays in the material that algorithms use to train their datasets. belong to more and more lawsuits.
“I expect a lot more to come,” said Robert DeBrowware, a digital media and intellectual property specialist at partner law firm Prior Cashman. He was not involved in the Silberman lawsuit.
