Last updated:
The world's first comprehensive horizontal legal framework for AI, establishing a risk-based classification system for systems deployed or used within the European Union.
This landmark regulation (EU 2024/1689) imposes tiered obligations based on risk: Prohibited, High-Risk, Limited Risk, and Minimal Risk. For high-risk systems, providers must implement a full Quality Management System (QMS), maintain technical documentation, and undergo conformity assessments. The Act affects any entity placing AI on the market or putting it into service in the EU, regardless of where the entity is headquartered, creating a global "Brussels Effect."
Real world example:
A San Francisco-based HR tech firm provides an AI-powered resume screening tool to a German client. Because "employment" is a high-risk category under Annex III, the US firm must produce technical documentation and appoint an EU authorized representative to comply with the EU AI Act.
