Source: Brussels Morning

Brussels Morning provides an updated analysis of the European Union AI Act as its various provisions become fully operational throughout 2025. The act is the first of its kind globally to provide a comprehensive, horizontal regulation of artificial intelligence based on a tiered risk hierarchy. It explicitly prohibits certain high-risk practices deemed incompatible with European values, such as social scoring by governments and untargeted scraping of facial images for recognition databases. For "high-risk" systems in sectors like infrastructure and law enforcement, the law mandates rigorous conformity assessments, human oversight, and the maintenance of high-quality training datasets to prevent algorithmic bias.

The article highlights the "Brussels Effect," observing how international companies are adjusting their global operations to match EU standards to maintain access to the single market. In 2025, the newly established EU AI Office has begun its active role in enforcing transparency for general-purpose AI (GPAI) models, requiring developers to disclose summaries of copyrighted content used for training. This regulatory environment is designed to foster trust among consumers while providing a predictable legal framework for businesses. However, the report notes that the success of the Act will depend on the effective coordination between national supervisors and the central EU authorities to ensure uniform application across all member states.