Hugging Face and Partners Outline Open ML Considerations for the EU AI Act

Hugging Face and a coalition of partners—including Creative Commons, Eleuther AI, GitHub, LAION, and Open Future—have issued a position paper outlining how the EU AI Act can better account for the needs of open, modular, and collaborative machine learning (ML) development. The goal is to ensure that the legislation supports robust innovation and transparency while achieving its regulatory objectives.

Key Recommendations for the EU AI Act

The coalition proposes five specific recommendations to ensure the EU AI Act remains compatible with the open-source ML ecosystem:

1. Clear Definition of AI Components

The EU AI Act should provide clear and unambiguous definitions of what constitutes an AI component. This clarity is essential for developers to understand their role and obligations under the law.

2. Protection for Collaborative Open Source Development

Building on the Parliament text’s Recitals 12a-c and Article 2(5e), the coalition recommends clarifying that the collaborative development of open-source AI components and their availability in public repositories does not automatically subject developers to the requirements of the AI Act.

3. Inclusive Governance and Coordination

The coalition advocates for the AI Office to maintain inclusive governance and coordination with the open-source ecosystem, following the guidelines set out in the Parliament's text.

4. Practical R&D Exceptions

To ensure research and development remain viable, the coalition proposes that the R&D exception be made practical and effective. This includes permitting limited testing in real-world conditions, combining elements of the Council’s approach and an amended version of the Parliament’s Article 2(5d).

5. Proportional Requirements for Foundation Models

Recognizing that different development modalities exist, the coalition recommends that requirements for "foundation models" (as detailed in the Parliament’s Article 28b) be proportional and distinctly treat open-source approaches differently from other uses.

The Role of Open ML in Innovation and Accountability

Open ML development is critical for creating robust innovation and supporting diverse, context-specific use cases. By allowing developers to share techniques and mix and match components with full visibility into the stack, open development fosters a more inclusive and accountable technological landscape. Hugging Face emphasizes that transparency—achieved through better documentation, accessibility of ML artifacts, and multidisciplinary collaborations—is instrumental in supporting the long-term goals of the EU AI Act.