A California court has partially dismissed a copyright case against OpenAI brought by several authors, including comedian Sarah Silverman, who allege OpenAI’s ChatGPT is pirating their work.
The case against OpenAI combines complaints filed by Silverman, Christopher Golden, Richard Kadrey, Paul Tremblay, and Mona Awad. (Awad left the suit in August.) It made six claims: direct copyright infringement; vicarious infringement; violation of the Digital Millennium Copyright Act (DMCA) by removing copyright management information; unfair competition; negligence; and unjust enrichment. OpenAI asked to dismiss all counts but the first and main complaint: direct copyright infringement.
The court ruled yesterday on OpenAI’s request to dismiss all but the direct infringement claim. In the ruling, Judge Araceli Martínez-Olguín threw out claims on vicarious copyright infringement, DMCA violations, negligence, and unjust enrichment. The court did not believe the plaintiffs’ allegations of unlawful business practices and fraudulent conduct related to unfair competition. It upheld the unfair competition claim that OpenAI did not seek their permission to use their work for commercial profit.
Judge Martínez-Olguín expressed skepticism of several of the authors’ claims. She wasn’t convinced of the allegation that OpenAI was intentionally removing copyright management information like the title and registration number, for instance, or that the authors had proven economic injury — since “nowhere in plaintiffs’ complaint do they allege that defendants reproduced and distributed copies of their books.” According to the court, the claim of “risk of future damage to intellectual property” was too speculative to consider. Martínez-Olguín also emphasized that the plaintiffs “have not alleged that the ChatGPT outputs contain direct copies of the copyrighted books” and “must show a substantial similarity between the outputs and the copyrighted materials.” The authors can file changes to their original complaint by March 13th.
While OpenAI won some concessions from the court, the main complaint that ChatGPT directly violated the authors’ copyrights remains on the table. Many of the other claims in the lawsuit hinge on proving direct infringement.
Tremblay first filed the suit in June, as reported by Reuters. Silverman’s complaint also listed Meta — through its large language model Llama 2 — as a defendant. The lawsuits alleged that OpenAI illegally copied their copyrighted work to train the large language model powering ChatGPT. If prompted to summarize the books written by the plaintiffs, they said ChatGPT generated accurate summaries, which they claim shows an intention to violate copyright.
OpenAI is facing several copyright infringement lawsuits from authors, including a proposed class action lawsuit from the Authors Guild and well-known authors like George R.R. Martin and John Grisham.