The EU publishes its first draft of a Code of Practice under the AI Act for general-purpose AI models, including provisions for copyright and systemic risk
A first draft of a Code of Practice that will apply to providers of general-purpose AI models under the European Union's AI Act has been published …
Context & Ripple Effects
The draft turns the AI Act from a finalized legal text into a more operational framework for general-purpose model providers. It follows lawmakers' earlier push for disclosure of copyrighted training material and the Act's publication in final form, which set the compliance path for companies.
Copyright and systemic-risk provisions put the focus on frontier-model governance rather than only on the Act's earlier high-risk-system framework. The later voluntary compliance code shows this implementation work continuing through provider guidance.
First-order effects
- Providers of general-purpose AI models gain an initial blueprint for documenting copyright-related practices and addressing systemic risks under the AI Act.
- The EU creates a common draft reference point for model developers, rights holders, and compliance teams as the code is refined.
Second-order effects
- Model providers may need to formalize training-data, documentation, and risk-management processes earlier, raising the importance of operational assurance capabilities.
- Copyright-focused provisions strengthen the practical relevance of the Act's earlier training-data disclosure direction, giving content owners a clearer regulatory channel for engagement with developers.
Third-order effects
- If codes of practice become the main route to implementation, AI regulation will increasingly be shaped through evolving technical documentation and risk-management standards alongside statutory rules.
- The approach could make governance expectations for broadly deployed models more comparable across providers, while leaving the eventual burden dependent on the final code and its application.
The trend: AI governance is moving from broad statutory principles toward provider-specific operating standards for powerful general-purpose models.