In a historic development, EU lawmakers have reached a political agreement on a comprehensive framework to regulate artificial intelligence (AI), marking a global first.
The deal, finalized after extensive negotiations, introduces a risk-based approach to AI regulation. It includes prohibitions on AI applications like biometric categorization based on sensitive characteristics, untargeted scraping for facial recognition databases, and emotion recognition in workplaces and educational institutions.
Strict safeguards limit law enforcement’s use of remote biometric identification, requiring judicial authorization for specific situations.
The legislation also addresses high-risk and general-purpose AIs and imposes obligations on AI systems that could potentially harm health, safety, or fundamental rights.
It introduces mandatory impact assessments for high-risk sectors, transparency requirements, and regulatory sandboxes for startups.
Penalties for non-compliance are significant, with fines up to €35 million or 7% of global turnover.
The law, which excludes military applications and non-professional AI use, aims to balance innovation with fundamental rights and European values.
Formal adoption of the final text is expected soon, with phased enforcement commencing as early as 2026.
Source: The Deep View