Penguin Random House amends its copyright notice globally to prohibit the use of books for training AI; the notice will be included in new titles and reprints
The world's biggest trade publisher has changed the wording on its copyright pages to help protect authors' intellectual property …
Context & Ripple Effects
Penguin Random House’s change makes an explicit reservation of AI-training rights part of the standard publishing workflow, rather than leaving the question to model developers’ interpretations of existing copyright pages.
The move sits between emerging licensing experiments, including HarperCollins’ reported opt-out AI-training arrangement, and a more contested legal environment, where the Anthropic ruling distinguished training from retaining pirated copies. It matters because publishers are beginning to formalize how books can enter AI data pipelines.
First-order effects
- New Penguin Random House titles and reprints will carry a clear notice opposing AI-training use, putting model developers and downstream data suppliers on notice of the publisher’s stated position.
- The notice creates a standardized rights-reservation record for the publisher and its authors, but it does not by itself set licensing terms or resolve whether a particular use is lawful.
Second-order effects
- Other publishers and literary rights holders face greater pressure to choose between comparable restrictions and negotiated access, rather than treating AI use as an unaddressed boilerplate issue.
- AI developers and dataset intermediaries will need more granular provenance and rights checks for books, especially where catalog-level permissions, exclusions, and author choices differ.
Third-order effects
- Book publishing is moving toward a rights-managed AI-input market in which traceable permissions may become more valuable than broad, poorly documented text collections.
- The eventual boundary will remain shaped by contracts and litigation: publishers’ allegations over Google’s book and article use show that notices can support a broader enforcement posture, but their independent legal force remains uncertain.
The trend: Copyright owners are converting AI-training objections into standardized rights controls while testing licensing as an alternative route to monetizing access.