Global AI rules diverge as EU phases in binding law

The European Union is the only major jurisdiction phasing in a comprehensive, binding AI law, while other governments rely on narrower measures, framework acts or existing regulators. Available regulatory developments largely extend through late 2025; scheduled changes could not be verified as of 14 October 2026, making official confirmation essential before compliance decisions.

The EU AI Act bans certain practices and imposes requirements based on risk. Most remaining obligations were scheduled for 2 August 2026, but a proposed Digital Omnibus could delay high-risk requirements. Whether that proposal was adopted remains unconfirmed. Penalties for banned practices reach €35 million or 7% of global annual turnover.

The US has no comprehensive federal AI statute, with policy driven mainly by executive orders and state legislation. California has advanced transparency, training-data disclosure and companion-chatbot rules, while other states regulate areas including hiring and high-risk systems. China instead uses targeted measures covering recommendations, deep synthesis, generative services and labeling of AI-generated content.

Japan has enacted a non-binding promotion framework without penalties. South Korea’s AI Basic Act was scheduled to take effect on 22 January 2026, covering high-impact systems and generative-output labeling. The UK, India, Canada and Türkiye have no AI-specific law in force under the latest verified information. Brazil’s risk-based bill passed the Senate but still requires lower-house approval.