Chicken Soup for the Soul is suing tech companies OpenAI, Anthropic, Google, Meta, xAI, Perplexity, Apple, and Nvidia for copyright infringement. The suit, filed March 17 in the Northern District of California, alleges that hundreds of its copyrighted works were ingested without authorization or compensation to train large language models.
In a noteworthy claim, the lawsuit alleges that in 2018, an OpenAI employee downloaded pirated books from Library Genesis—a shadow library repeatedly enjoined by federal courts—to create internal datasets later used to train GPT-3. That decision, the filing argues, set an example for every other defendant to follow.
Within weeks of GPT-3's release, the open-source collective EleutherAI assembled "Books3," approximately 200,000 books copied from Bibliotik, a private BitTorrent tracker, and packaged it into a publicly distributed dataset called The Pile. The remaining defendants, the complaint alleges, then adopted The Pile as a foundational training resource, with some expanding to additional shadow libraries.
"Each escalation compounded the last: from a single LibGen download, to an industry-standard pirated dataset, to the wholesale harvesting of every major shadow library on the internet," the filing states.
The plaintiff characterizes the conduct as "not isolated acts of infringement, but an industry-wide course of conduct." Infringement, the complaint adds, did not end with the initial downloads—defendants reproduced copyrighted works "potentially countless" additional times through preprocessing, deduplication, and iterative model fine-tuning.
The complaint also argues that the Chicken Soup for the Soul series was intentionally targeted for its "tightly edited, first-person narratives written in natural, conversational language that conveys emotion, moral reflection, and coherent storytelling in concise form"—characteristics that made the franchise, in the plaintiffs' framing, "uniquely well suited for training large language models to replicate authentic human voice, narrative pacing, emotional tone, and story structure."
Much like the complaint filed in December by author John Carreyrou and others against many of the same defendants, this filing also aims to challenge the class-action model that has dominated AI copyright litigation.
Pointing to the pending Anthropic settlement in the Northern District of California, the suit notes that the framework would pay rights holders approximately $3,000 per work—"just 2% of the Copyright Act's statutory ceiling of $150,000 per willfully infringed work." The complaint states that such settlements "seem to serve Defendants, not creators."
Chicken Soup for the Soul is instead seeking individualized statutory damages determined by a jury. The law firms behind the suit say more than 1,000 authors representing more than 5,000 works have signed on to the same approach.
