Publisher Wiley tells AI developers: Get a license before training on our content

John Wiley & Sons has released a formal position statement on “illegal scraping of copyrighted content by AI developers.” The notice makes it explicit that training or operating artificial-intelligence systems on Wiley-owned literature requires a licence, countering any suggestion of implied permission.

The publisher’s stance lands amid an industry-wide debate over the provenance of large-language-model (LLM) training data. While acknowledging “the transformative potential of AI to advance research and discovery,” Wiley stresses that progress must rest on respect for authors’ intellectual-property rights.

In a single sweep, the statement (i) insists there is no implied permission to ingest its catalogue, (ii) positions data provenance and clear attribution as non-negotiable pillars of ethical AI practice, and (iii) points to an active licensing market. Wiley’s Q3 FY 2025 results show $9 million in AI-training fees booked in that quarter alone.

Wiley frames this licensing approach as ultimately beneficial not just for creators but also for the AI systems themselves and their users. By ensuring AI is grounded in authoritative, high-quality content through transparent agreements, including expectations for citation and attribution, the publisher believes greater user trust can be fostered.

A recent example of Wiley’s proactive, licensed approach is its involvement in the European Space Agency’s (ESA) “EVE” (Earth Virtual Expert) project, developed with Italy-based AI specialist Pi School. Instead of relying on broad web scraping, EVE will be trained on a specifically licensed, curated dataset from Wiley, including 71 peer-reviewed journals and other authoritative Earth-science materials.

Complementing its stance on external AI development, Wiley has also issued detailed guidelines for its authors using generative AI. In a nutshell, these guidelines permit the use of AI tools as writing companions but stress that human oversight, verification of accuracy, and disclosure of AI use to Wiley are mandatory. Authors are explicitly warned to review AI tool terms and conditions carefully to avoid granting unintended rights (like permission to train on their content) to the AI provider. Ultimately, authors retain full responsibility for their work’s originality, accuracy, and ethical integrity.