European UnionAI regulation

Regulation Status

• EU AI Act is the central, binding regulation for AI, applying a risk-based approach across all sectors, including rules for general-purpose AI models and those posing systemic risk. Implementation remains phased, with prohibitions on unacceptable risk systems, GPAI model obligations, and applicable transparency and enforcement provisions already in force; from 2 August 2026, the AI Office, in collaboration with national authorities, began enforcing applicable AI Act rules including transparency duties for certain AI systems such as disclosure of AI interactions and synthetic or altered content. Digital Omnibus amendments approved by the European Parliament on 16 June 2026 and adopted by the Council of the European Union on 29 June 2026 delay high-risk obligations to 2 December 2027 for stand-alone high-risk AI systems and 2 August 2028 for embedded systems covered by EU sectoral safety legislation, delay AI-generated-content watermarking obligations to 2 December 2026, centralise aspects of enforcement through the AI Office, reduce regulatory overlap, extend certain SME exemptions to small mid-cap enterprises, expand access to AI regulatory sandboxes, and add prohibitions on AI systems generating child sexual abuse material or non-consensual intimate depictions without adequate safeguards.
• Ongoing consultations, amendments, and sector-specific initiatives continue in areas including health, finance, copyright, biotech, cloud and compute infrastructure, and child protection, with strong interaction between the AI Act, GDPR, competition law, product safety and liability rules, and international cooperation frameworks.
• Numerous guidelines, codes of practice, and administrative guidance supplement the AI Act, clarifying definitions, prohibited practices, GPAI model obligations, transparency, copyright, high-risk system classification, serious incident reporting, and marking or labelling of AI-generated content. The AI Liability Directive was withdrawn in February 2025 and is not expected to be revived; liability for AI is now addressed through the revised Product Liability Directive and existing national tort law.