Judge dismisses most of Silverman's book-author lawsuit against OpenAI

US District Judge Araceli Martinez-Olguín dismissed most of the claims brought by Sarah Silverman, Michael Chabon, Ta-Nehisi Coates and other authors against OpenAI, in a case that had followed the original filing the previous July. The ruling dismissed claims of vicarious copyright infringement, violation of the Digital Millennium Copyright Act, negligence and unjust enrichment.

On vicarious infringement, the central issue was that the authors had not adequately alleged that ChatGPT’s outputs were themselves substantially similar to their copyrighted books — the complaint’s evidence, drawn from prompting the model to summarise the plaintiffs’ work, showed the model could describe the books but did not demonstrate output closely resembling the original text. The judge gave the plaintiffs until mid-March to file an amended complaint addressing that gap on the dismissed counts.

What survived was the claim at the centre of the suit: that OpenAI’s act of copying the authors’ books into its training data was itself direct copyright infringement, regardless of what ChatGPT later output. A California unfair-competition claim tied to that same underlying conduct also proceeded. That distinction — between infringement in the training process and infringement in the model’s output — became one of the central doctrinal fault lines running through the wider run of author lawsuits against AI developers, with courts in parallel cases reaching different views on how much resemblance a plaintiff needed to show and whether training itself, independent of output, was actionable.