The Lawfare Podcast Lawfare Archive: Itsiq Benizri on the EU AI Act
12 snips
Jul 5, 2025 Itsiq Benizri, a counsel at WilmerHale Brussels, dives into the recently agreed EU AI Act and its global implications. They discuss how domestic politics influenced the final legislation and the significance of its risk-based classification of AI systems. Benizri highlights the challenges in implementing the Act and the balance between fostering innovation and setting minimum standards. The conversation also touches on the regulatory approach to deep fakes, emphasizing the need for transparency amidst the rapidly evolving AI landscape.
AI Snips
Chapters
Transcript
Episode notes
EU AI Act Risk-Based Approach
- The EU AI Act uses a risk-based approach, classifying AI into four levels from minimal risk to unacceptable risk with varying requirements.
- It applies to providers, importers, distributors, manufacturers, and deployers, but excludes military and some open-source AI.
Rushed Finalization Challenges
- Political agreement reached in December 2023 outlined key principles; finalizing technical text took additional months.
- The rushed process led to legal drafting quality not meeting EU standards, posing implementation challenges.
Late Negotiation Uncertainty Causes
- Uncertainty late in negotiations stemmed from significant issue importance, divergent views among lawmakers, and looming European elections.
- Debates included how to regulate facial recognition and generative AI, reflecting political and national tensions.