OpenAI moves to dismiss most claims in authors' ChatGPT training-data class actions

What happened

OpenAI has filed a motion to dismiss most of the claims in two near-identical class-action complaints brought by authors over ChatGPT training data. One suit is led by Sarah Silverman (with Christopher Golden and Richard Kadrey); the other by Paul Tremblay, author of *The Cabin at the End of the World*. Both allege OpenAI copied books into training datasets "without consent, without credit and without compensation." OpenAI asks the San Francisco federal court to dismiss claims for vicarious copyright infringement, DMCA violations, unfair competition, "negligence," and unjust enrichment. The motion does NOT seek dismissal of the direct copyright infringement claim, which OpenAI says it will "seek to resolve as a matter of law at a later stage." OpenAI is represented by Morrison & Foerster and Latham & Watkins; plaintiffs by Joseph Saveri and Matthew Butterick.

Why it matters to creators

These are among the highest-profile author class actions against OpenAI. Which claims survive a motion to dismiss determines the legal theories creators can use going forward — the motion argues, for example, that DMCA and unfair-competition claims do not satisfy statutory elements, and that ChatGPT outputs are not necessarily infringing "derivative works" merely because they were "based on" a dataset that included plaintiffs' books. A ruling here will shape how future training-data cases are pleaded.

What is not yet known

The court has not ruled on the motion; the direct-infringement claim remains pending regardless. A motion to dismiss narrows claims, it does not decide liability.