Understanding the Latest Developments
A new draft of the Code of Practice for general purpose AI (GPAI) models has been released ahead of a May deadline for finalizing guidance under the EU AI Act. This draft aims to clarify compliance requirements for AI model makers, helping them avoid penalties that could reach up to 3% of their global annual revenue. The Code has been revised to be more streamlined, incorporating feedback from previous drafts to enhance clarity and coherence.
Key Highlights of the Draft
- The draft includes sections on commitments for GPAIs, transparency, copyright, and safety obligations.
- It offers guidance on transparency, including a model documentation form for compliance.
- Copyright remains a contentious issue, with vague terms like “best efforts” allowing AI companies some leeway in data acquisition.
- The safety and security obligations have been further narrowed, focusing only on the most powerful models.
Why This Matters
The ongoing development of the Code reflects the EU’s efforts to regulate AI while facing pressure from the U.S. to ease restrictions. As AI technology evolves, so do the legal frameworks surrounding it. The outcome of this draft will have significant implications for how AI companies operate within Europe and could set a precedent for global AI regulations. The balance between innovation and regulation is crucial, as overregulation may stifle growth in this rapidly advancing field.
