The European Commission has taken a key step toward the practical implementation of the Artificial Intelligence Act (AI Act) by officially publishing draft guidelines for classifying high-risk AI systems. This document aims to provide legal clarity for developers and companies using advanced technologies in the EU.
The main goal of the new guidelines is to clearly delineate technologies that may pose a direct threat to the safety, health, and fundamental rights of citizens. The “high-risk” category, which requires strict oversight and pre-certification before entering the EU market, includes AI solutions in the following areas:
Biometric identification: Facial and emotion recognition systems in public spaces.
Critical infrastructure: AI managing traffic, water supply, or electricity grids.
Human resources (HR): Algorithms for automated resume screening, candidate assessment, and employee monitoring in the workplace. Justice and Migration Control: Tools for Predicting Offenses and Assessing Risks in Visa Issuance.
“This document translates the theoretical provisions of the AI Law into clear and binding rules. We are creating a secure digital environment without hindering technological progress,” Brussels emphasizes.
The publication of the draft opens the final consultation phase. Public comment on the guidelines will last until June 23, 2026. During this month, representatives of the IT industry, human rights activists, and experts can submit their amendments and comments.
Once the final version of the document is approved, these requirements will become strictly mandatory for all AI developers operating in the European market. Companies that fail to comply with the classification and security standards will face multi-million-dollar fines.