Stand-up comedian Sarah Silverman has initiated legal action against OpenAI and Meta, alleging that their AI models unlawfully utilized her content for training purposes without her consent. Silverman, along with authors Christopher Golden and Richard Kadrey, claim that both OpenAI’s ChatGPT and Meta’s LLaMA were trained on unlawfully-acquired datasets that contained their works. The authors assert that these AI models utilized “shadow library” websites, such as Bibliotik, Library Genesis, and Z-Library, where most of the content uploaded is protected by authors’ intellectual property rights. When asked to create a dataset, ChatGPT produced a list of titles sourced from these illegal online libraries. The legal complaint, filed in San Francisco federal court, also includes another suit against Meta Platforms, the parent company of Facebook.
The suit includes exhibits that demonstrate ChatGPT’s responses when asked to summarize books authored by Silverman, Golden, and Kadrey. However, the synopses generated by ChatGPT failed to reproduce any copyright management information included by the authors in their published works, despite providing very accurate summaries. This suggests that ChatGPT retains knowledge of specific works in its training dataset and is capable of generating similar textual content. The authors’ suit against Meta also highlights the alleged use of illicit sites to train LLaMA, a competitor to ChatGPT launched by Mark Zuckerberg’s company.
All AI models are trained using extensive datasets and algorithms. LLaMA, for instance, utilizes a dataset called The Pile, which was assembled by EleutherAI, a nonprofit AI research group. The authors’ complaint points to a paper published by EleutherAI that details the derivation of one of its datasets, called Books3, from a copy of the contents of the Bibliotik private tracker. The lawsuit argues that both ChatGPT and LLaMA used datasets for training that sourced content from illicit “shadow libraries” mentioned within the legal documents, such as Bibliotik.
The authors maintain that they did not provide consent for their copyrighted books to be used as training material for either AI model. According to the claims, OpenAI and Meta have therefore violated several copyright laws, including negligence, unjust enrichment, and unfair competition. While the damage caused by these violations “cannot be fully compensated or measured in money,” the plaintiffs are seeking statutory damages, restitution of profits, and more.
The legal counsel representing the authors, Joseph Saveri and Matthew Butterick, are known for their involvement in multiple lawsuits relating to authors and AI models. In 2022, they filed a lawsuit against OpenAI’s GitHub Copilot, which converts natural language into code and was acquired by Microsoft for $7.5 billion in 2018. They claimed violations of privacy, unjust enrichment, unfair competition laws, and fraud, among other allegations. Saveri and Butterick have also challenged AI image generator Stable Diffusion earlier this year and have represented numerous book authors in class-action litigation against AI technology.
As of now, the authors’ legal counsel has not responded to The Post’s request for comment. The Post has also reached out to OpenAI and Meta for their statements on the matter.
Denial of responsibility! VigourTimes is an automatic aggregator of Global media. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, and all materials to their authors. For any complaint, please reach us at – [email protected]. We will take necessary action within 24 hours.
