Source: The Lens - Slaughter and May

Richard Barker, Senior Knowledge Lawyer at Slaughter and May, provides a detailed examination of the copyright chapter in the final version of the EU’s General-Purpose AI (GPAI) Code of Practice published on 10 July 2025. The chapter contains five key measures that GPAI model providers agree to implement to demonstrate compliance with Article 53(1)(c) of the AI Act.

The measures include: (1) drafting and maintaining a copyright policy; (2) ensuring lawful reproduction and extraction of web-crawled content without circumventing technological protection measures; (3) identifying and complying with rights holders’ text and data mining (TDM) opt-outs, including adherence to robots.txt and other machine-readable protocols; (4) mitigating risks of copyright-infringing AI outputs through technical safeguards and acceptable use policies; and (5) designating a point of contact and complaint mechanism for rights holders.

The final version of the copyright chapter has seen some material changes, including the removal of a due diligence measure related to non-web-crawled content and stronger commitments from signatories. However, the chapter remains voluntary and does not guarantee compliance with EU copyright law, which is separately enforced.

The article highlights ongoing challenges, such as the complexity of determining lawful use of web content and the need for further guidance. It also notes the Commission’s FAQs that encourage good faith adherence and collaboration rather than immediate penalties for partial compliance.