EU: AI Act Requires Member States to Establish AI Regulatory Sandboxes Amid Implementation Challenges
On March 17, 2026, the European Parliamentary Research Service (EPRS) published a briefing regarding the state of play and the implementation challenges of artificial intelligence (AI) regulatory sandboxes under the EU Artificial Intelligence Act (AI Act).
The AI Act mandates that Member States establish or participate in at least one AI regulatory sandbox, which is designed to enable the supervised testing of AI systems for compliance with the regulation.
Understanding AI Regulatory Sandboxes
The introduction of AI regulatory sandboxes serves as a vital tool to support innovation while ensuring adherence to EU AI rules. These sandboxes are controlled environments where AI systems can be developed, trained, validated, and tested under the supervision of national authorities.
Key Takeaways
- AI Regulatory Sandboxes: The AI Act requires Member States to create or join at least one AI regulatory sandbox by August 2, 2026.
- Purpose: These sandboxes facilitate supervised testing of AI systems to confirm compliance with EU AI regulations.
- Implementation Challenges: Notable challenges encompass design, fragmentation, and timing issues.
- Coordination: National authorities must synergize with EU bodies, including the European Commission and the European AI Board.
Things to Consider
Member States should proactively address the identified implementation challenges related to design, fragmentation, and timing to ensure the effective deployment of AI regulatory sandboxes.
