Lawsuit tracker
Authors v. Apple
Hendrix v. Apple Inc., No. 25-cv-07558 (N.D. Cal.), consolidated with Martinez-Conde v. Apple and Alexander v. Apple
| Plaintiffs | Authors Grady Hendrix and Jennifer Roberson, joined in the consolidated complaint by John Hornor Jacobs and Eboni McKinnon, on behalf of a proposed class |
|---|---|
| Defendants | Apple Inc. |
| Court | U.S. District Court for the Northern District of California (Judge Yvonne Gonzalez Rogers) |
| Filed | 2025-09-05 |
| Status | Active |
| Content type | Books — fiction and nonfiction allegedly included in the pirated Books3 dataset |
| Last updated | 2026-07-21 |
The claims
Copyright infringement: alleged use of the pirated Books3 dataset to train Apple’s OpenELM models and alleged use of shadow-library books in training Apple Intelligence foundation models
What has happened
Novelists Grady Hendrix and Jennifer Roberson sued Apple in September 2025, alleging it trained its OpenELM language models on Books3, a dataset of nearly 200,000 pirated books, and concealed the sources of training data behind Apple Intelligence. The court consolidated the case with two similar author suits, Martinez-Conde and Alexander, in November 2025, and appointed interim co-lead class counsel from Keller Rohrback and Susman Godfrey in January 2026. A consolidated class action complaint followed on February 13, 2026, adding authors John Hornor Jacobs and Eboni McKinnon. Apple answered in spring 2026 rather than moving to dismiss, denying the claims and contending that any use of the works was fair use. The case is in its early merits phase.
Key developments
- 2025-09-05 — Hendrix and Roberson file a proposed class action alleging Apple trained OpenELM on the pirated Books3 dataset.
- 2025-11-14 — The court consolidates Hendrix with the Martinez-Conde and Alexander author suits.
- 2026-01-30 — Interim co-lead class counsel appointed from Keller Rohrback and Susman Godfrey.
- 2026-02-13 — Plaintiffs file a consolidated class action complaint on behalf of Hendrix, Roberson, John Hornor Jacobs and Eboni McKinnon.
- 2026-04-30 — Apple files an answer to the operative complaint, denying the claims and asserting fair use.
Why it matters for training data
Apple shows that a small research model trained on a public pirated dataset carries the same class-action tail as a frontier model, years after the training run. Apple chose to answer and litigate fair use rather than seek early dismissal, a sign these cases now go to the merits. For buyers, anything touching Books3 or The Pile needs provenance documentation; for suppliers, a dataset being taken offline does not erase downstream exposure.
Sources
- CNBC: Apple sued by authors over use of books in AI training
- Keller Rohrback: Apple Intelligence AI copyright litigation case page
- Chat GPT Is Eating the World: Hendrix v. Apple filing analysis
- CourtListener docket: Hendrix v. Apple Inc.
- Apple’s answer asserting fair use (filed April 30, 2026)
Deeper analysis
Want data that clears this in diligence?
Whether you're building a model or sitting on an archive, the first conversation is short and specific.
Send a brief