The EU AI Act
The European Union's AI Act takes a risk-based approach: AI systems are classified as unacceptable risk (prohibited), high risk (subject to conformity assessment before deployment), limited risk (transparency requirements) or minimal risk (largely unrestricted). High-risk categories include AI in critical infrastructure, education, employment, essential services, law enforcement and administration of justice. Generative AI systems face transparency requirements including disclosure that content is AI-generated.
The US approach
The US has not passed comprehensive federal AI legislation. The approach relies on existing regulatory bodies — the FDA for medical AI, the FTC for consumer AI harms, sector-specific financial regulators — supplemented by executive orders and voluntary commitments from major AI companies.
China
China has enacted regulations targeting specific AI applications: algorithm recommendation systems must not create filter bubbles that undermine social cohesion; generative AI services must ensure content reflects core socialist values. The regulatory focus is as much on political and social alignment as on safety in the conventional sense.