Sarah Silverman, along with authors Richard Kadrey and Christopher Golden, has filed a lawsuit against Meta Platforms’ LLaMa and OpenAI’s ChatGPT, accusing them of copyright infringement. The plaintiffs claim that Meta and OpenAI used their copyrighted content without permission to train their respective AI systems.
According to court documents, Meta admitted to using the plaintiffs’ copyrighted books in the dataset to train LLaMa. Similarly, the lawsuit against OpenAI alleges that ChatGPT’s summaries of the plaintiffs’ work indicate training with copyrighted content, despite some inaccuracies.
The lawsuits assert that Meta and OpenAI obtained the copyrighted data from “shadow libraries,” such as Bibliotik, Library Genesis, and Z-Library, which distribute copyrighted material through torrent systems. These shadow libraries contain large quantities of copyrighted material, making them attractive to the AI training community.
In addition to seeking remedies for the infringement of their personal work, the authors are filing the complaint on behalf of a class of copyright owners across the United States whose works were allegedly infringed.
Both Meta and OpenAI were contacted for comment on the case but did not respond prior to publication. This lawsuit follows a recent strike by writers in the Writers Guild of America, which brought attention to various issues in the industry, including the use of AI.
This legal action highlights the importance of addressing copyright concerns in AI development, especially when it comes to using copyrighted materials without appropriate permissions. It also raises questions about the responsibility of AI developers to ensure compliance with intellectual property rights.
The outcome of this lawsuit could have implications for the future use of copyrighted content in AI training and may lead to increased scrutiny and regulations surrounding the sourcing of training data.