European publishers push for tougher AI copyright enforcement

The Federation of European Publishers welcomed the European Parliament’s adoption of a non-legislative report on AI and copyright, saying it acknowledges longstanding concerns from the book sector about generative AI training practices. The group argues that AI developers have relied on copyright-protected content without respecting European copyright law, while AI-generated material increasingly competes with creative works.

FEP says US court cases have shown that books are considered especially valuable training data for generative AI systems, and that some developers used illegal sources rather than working with publishers. The organization also points to a rise in “fake books”, described as AI-generated texts that do not disclose their nature to consumers and have spread across online book marketplaces.

The publishers’ group supports the report’s call for stronger transparency requirements, describing them as essential for licensing and enforcement. It also backs measures such as a presumption of use when AI companies fail to disclose training data, while stressing that rightsholders must retain the ability to refuse licences and should not face mandatory or one-size-fits-all licensing schemes.

FEP President Sonia Draga called on the European Commission and Member States to enforce existing law and provide publishers with stronger tools to defend their rights against generative AI developers accused of ignoring Europe’s rules.