A new study by Yann Padova and Sebastian Thess, commissioned by the Computer & Communications Industry Association (CCIA Europe), was released today: "The AIA Code of Practice – Opportunities & Challenges."

The study offers concrete recommendations for designing the AI Act’s upcoming code of practice for providers of general-purpose AI (GPAI) models that will be critical for the success of the European AI ecosystem. It seeks to inform the code’s drafting process, which the European Commission will kick off in the second half of September.

As a rapidly evolving frontier technology, GPAI holds immense potential for society. A poorly drafted GPAI code of practice, on the other hand, could put a brake on innovation in the European Union – making it essential for this code to succeed.

The study identifies key factors for the code’s success by drawing on the EU’s past (co-)regulatory experiences. The authors also emphasize the importance of aligning the EU framework with international emerging approaches and standards, ensuring legal clarity, and following an effective drafting process.

GPAI providers have to play a key role in the drafting of the code, according to Padova and Thess, given their unique expertise and position as the primary subjects of these rules. The study also suggests focusing on operationalizing the rules introduced by the AI Act while explicitly avoiding discussions on topics outside the Act’s scope, such as the EU copyright framework.