The European Union AI Act, in its rightful name Regulation (EU) 2024/1689 of 13 June 2024 laying down harmonised rules on artificial intelligence and amending [several other regulations] (Artificial Intelligence Act), was published in the Official Journal of the EU on 12 July 2024 (see here). However, for an easier reading of the AI Act I recommend the AI Act Explorer webpage published by the Future of Life Institute.
The AI Act It is the first legislation to regulate AI systems, and aims to provide a framework for the development, market placement, and use of AI systems. It classifies AI systems into four risk levels:
- Unacceptable Risk: Prohibited practices that violate EU values and rights.
- High Risk: AI systems that significantly impact safety or fundamental rights, requiring stringent requirements.
- Specific Transparency Risk: Systems with manipulation risks, necessitating transparency.
- Minimal Risk: Most AI systems with no specific obligations.
General-purpose AI models, like large language models, have specific frameworks and obligations due to their broad applicability and potential risks.
Implementation Timeline
- 1 August 2024: AI Act enters into force.
- 2 February 2025: Prohibitions on unacceptable risk AI systems take effect, as well as AI literacy requirements (I recommend Article 3.56 to see what that means).
- 2 August 2025: Rules for general-purpose AI models and designation of national authorities come into force.
- 2 August 2026: Full application of high-risk AI system regulations and establishment of AI regulatory sandboxes.
- 2 August 2027: Rules for high-risk AI systems in specific sectors, like toys and medical devices, apply.