European Commission Unveils Landmark Risk-Based AI Regulation
The European Commission publishes a comprehensive proposal for an Artificial Intelligence Act that bans harmful practices like social scoring while imposing strict obligations on high-risk systems.
The European Commission publishes a landmark Proposal for an Artificial Intelligence Act to establish harmonised rules across the bloc. This 108-page regulation builds on previous EU strategies and relies on a risk-based approach to govern artificial intelligence. It defines an AI system as software developed using techniques like machine learning that generates outputs influencing the environments they interact with.
Under the new proposal, the EU strictly prohibits AI practices that pose unacceptable risks to society. The ban targets systems that materially distort human behaviour and cause physical or psychological harm, government-led social scoring, and real-time remote biometric identification in public spaces. Law enforcement agencies face narrow exceptions for the biometric ban, limited to prosecuting severe crimes or preventing imminent threats to life and terrorist attacks.
The act also imposes strict obligations on high-risk AI systems outlined in the regulation. These high-risk categories include AI applications deployed in education, employment, and worker management. By separating AI systems into prohibited, high-risk, and acceptable categories, the European Commission aims to foster trustworthy AI innovation while protecting fundamental rights.