Overview
Severity: MEDIUM | Affected: European Union | Category: policy
The European Parliament has formally ratified the final technical provisions of the landmark EU AI Act, with a key new requirement for mandatory third-party audits for all 'high-risk' AI systems. Effective immediately for new systems, providers of AI in critical sectors like healthcare, finance, and critical infrastructure must undergo rigorous security and bias audits by accredited independent bodies before their products can be marketed in the EU. These audits will assess systems for robustness against adversarial attacks, data privacy compliance, and algorithmic fairness. The move is a significant step towards establishing a concrete enforcement mechanism for AI safety and ethics. Critics, however, express concern over the limited number of accredited auditing firms and the potential for the process to slow down innovation.