June 29 (Reuters) – Two U.S. authors sued OpenAI in San Francisco federal court on Wednesday, accusing the company of misusing its work to “train” its popular generative artificial intelligence system ChatGPT, in a proposed class action lawsuit. claimed to have done so.
Massachusetts-based authors Paul Tremblay and Mona Awad said ChatGPT infringes authors’ copyrights by mining data copied without permission from thousands of books.
Matthew Buterick, an attorney for the authors, declined to comment. Representatives for OpenAI, a privately held company backed by Microsoft (MSFT.O), did not respond to a request for comment.
Several legal challenges have been filed over materials used to train cutting-edge AI systems. Plaintiffs include source code owners against OpenAI and Microsoft’s GitHub, and visual artists against Stability AI, Midjourney, and DeviantArt.
The plaintiffs allege that their system makes fair use of copyrighted works.
ChatGPT provides conversational responses to user text prompts. It became the fastest growing consumer application of all time earlier this year, reaching his 100 million active users in January, just two months after its launch.
ChatGPT and other generative AI systems use large amounts of data collected from the internet to create content. The Tremblay and Awad lawsuit said books were “a key element” because they provided “the best examples of high-quality long-form writing.”
OpenAI’s training data is estimated to have incorporated more than 300,000 books, according to the complaint, including books from illegal “shadow libraries” that offer copyrighted books without permission. .
Awad is best known for his novels “13 Ways to See Fat Girls” and “Bunny”. Among Tremblay’s novels is The Cabin at the End of the World, which was adapted in M. Night Shyamalan’s movie Knock at the Cabin, which was released in February.
Tremblay and Awad said ChatGPT can generate “very accurate” summaries of their books, showing that they’re in the database.
The lawsuit seeks an unspecified amount of damages on behalf of a nationwide copyright owner whose work OpenAI allegedly misused.
Reporting by Blake Britten, Washington Editing by David Barrio and Richard Chan
Our standards: Thomson Reuters Trust Principles.
