The first draft of the General-Purpose AI Code of Practice is published

← All articles
Artificial intelligence · EN

The first draft of the General-Purpose AI Code of Practice is published

This initial version, prepared by Chairs and Vice-Chairs of four thematic working groups, sets the foundation for guiding the development and deployment of trustworthy and safe general-purpose AI models.

Next week, nearly 1000 stakeholders, including EU Member States representatives and international observers, will discuss this draft in dedicated working group meetings. Feedback from these discussions will help shape the future iterations of the Code, aiming for a comprehensive final version by 1 May 2025.

Key highlights:
- The draft outlines guiding principles and objectives for the Code.
- It includes transparency and copyright-related rules for AI model providers.
- It addresses systemic risks and proposes risk assessment and mitigation measures.

Remarkable Aspects:
1. Comprehensive Input: The draft is informed by nearly 430 stakeholder submissions, international best practices, and key frameworks like the G7 Code of Conduct and the AI Act.
2. Guiding Principles: It is built on six key considerations: alignment with Union principles, proportionality to risks and provider size, support for AI safety, and future-proofing.
3. Collaborative Approach: The draft is open to feedback, aiming to refine Measures, Sub-Measures, and KPIs with input from diverse perspectives.
4. Focus on Systemic Risks: It acknowledges the potential for a small number of AI models to pose systemic risks, suggesting a tiered system of measures if needed.

Stakeholders are invited to provide feedback through the Futurium platform by 28 November, 12:00 CET.

This post has a document attached on LinkedIn. View on LinkedIn →

Discuss on LinkedIn