EU Proposes Sweeping Risk-Based AI Regulation With Strict Bans
The European Commission unveils a comprehensive AI regulation that uses a risk-based approach to govern artificial intelligence across various sectors. The proposal prohibits four specific harmful AI practices, including subliminal manipulation and social scoring by public authorities.
The European Commission unveils a highly anticipated proposal that establishes harmonized rules for artificial intelligence across the continent. This regulation applies broadly to providers and users of AI systems located within the EU, as well as foreign entities whose AI outputs are utilized within the union. The commission defines an AI system broadly as software developed using specific techniques that generates content, predictions, recommendations, or decisions influencing its environment.
Lawmakers take a cautious, risk-based approach to the technology by balancing its potential benefits against the threats it poses to European values and fundamental rights. To protect citizens, the proposal explicitly prohibits four categories of harmful AI practices. These bans target systems that use subliminal techniques to distort behavior and cause harm, as well as AI that exploits the vulnerabilities of specific groups based on age or disability.
Additionally, the regulation forbids public authorities from using AI systems to evaluate or classify the trustworthiness of individuals through social scoring. This comprehensive framework serves as a foundational piece of the broader European Strategy for data. By setting these strict boundaries, the EU aims to foster trustworthy AI development while shielding the public from manipulative and dangerous applications of the technology.