Share This Article
Negotiations on the EU AI Act are becoming intense and the main subject of debates relates to the introduction on regulations on foundation models.
We are at a pivotal moment for the EU AI Act, the Spanish Secretary of State for Digitalisation and Artificial Intelligence, Carme Artigas, offered crucial insights summarized below:
Foundation models: Recent negotiations encountered a hurdle regarding these essential models, key to AI systems like ChatGPT. Concerns from France, Italy, and Germany prompted a reevaluation.
Balancing Act: Artigas underscores the need to address foundation models’ systemic risks without overly restricting the technology-neutral AI law.
Transparency Focus: Instead of high-risk categorization, the approach now centers on refining the legal text with codes of conduct and transparency obligations. However, stricter obligations for “high-impact” foundation models and a compromise proposal are still under discussion. Talks are set for November 21st.
Copyright Considerations: Discussions also revolve around handling copyrighted content, striking a balance between legal compliance and European competitiveness.
Adaptive Regulation: Artigas emphasizes the importance of flexible regulation to adapt to technology’s evolution and ensure accountability and enforceability.
Artigas acknowledges the fast-paced AI landscape, urging Europe to remain proactive amid global advancements. There is no doubt that Spain is pushing to have the AI Act finalized by the end of the year during their EU Presidency.
We need to follow closely the coming weeks as they will be crucial for the future of artificial intelligence in Europe. Lack of clear regulations are unlikely to set a precise regime which shall foster investments within the European Union with the goal of making it become a global standard for other jurisdictions.
On a similar topic, you may find the following article interesting “New Obligations for Foundation Models and General Purpose AI Systems”.
