EUResearch
SSRN: Llopis, Prastacos and Vonortas propose review-board AI regulation modeled on IRB and FDA precedent
In an October 10, 2026 SSRN preprint, Antonio Llopis, Gregory Prastacos and Nicholas Vonortas argue that AI regulation should center expert review boards under ethical principles that evolve by precedent, applying that Review-Board-Based framework to the EU AI Act's risk tiers by replacing fixed statutory classifications with evolving expert review and extending Belmont-style protections to AI-affected communities.
Key points
- Llopis, Prastacos and Vonortas posted the preprint on SSRN on October 10, 2026.SSRN
- They argue open-ended AI risks fit expert review boards under ethical principles that evolve by precedent.SSRN
- Their Review-Board-Based framework is designed as an input to risk-based regimes such as the EU AI Act.SSRN
- They would keep the Act's risk tiers but replace fixed classifications with expert-review precedent evolution.SSRN
- They extend Belmont principles to users, third parties and populations affected by systemic AI deployments.SSRN
- A worked case study illustrates the classification mechanism and community-protection reformulation.SSRN