EU AI Act sets risk-based rules with phased deadlines through 2028

The European Commission proposed an AI regulation in 2021.[2] The European Parliament approved the final text in March 2024, and EU governments in the Council backed it two months later, Meer reports.[1] The final version appeared in the EU’s Official Journal on 12 July 2024, and the Act entered into force on 1 August that year.[2][3] Entry into force did not mean all provisions applied at once: the Commission’s timeline schedules full implementation by 2 August 2028.[3] The Council and Parliament adopted Regulation (EU) 2026/1744, known as the Digital Omnibus on AI, on 29 May 2026; the evidence provided does not explain how it affects the timetable.[2]

The Act assesses AI systems according to their risks and general-purpose AI models according to their capabilities.[2] It bans systems classed as presenting “unacceptable risk,” with social scoring and unauthorized real-time biometric surveillance among the practices Meer identifies as prohibited.[1] High-risk uses, including AI in healthcare, policing and hiring, are allowed only under strict conditions.[1] The Act requires those systems to meet standards for transparency, oversight and safety testing.[1] For general-purpose AI, the Commission’s timeline says related obligations and governance requirements began applying on 2 August 2025.[3]

The Commission says certain AI systems must tell people when they are interacting with AI and when content has been generated or altered by AI.[4] Chatbots must identify themselves as AI rather than a human, and deepfakes must be labelled.[4] The Commission says those transparency measures are intended to reduce deception and manipulation and help people make informed choices.[4] Its timeline says most rules and enforcement for applicable provisions began on 2 August 2026.[3] Rules for high-risk systems in Annex III are scheduled for 2 December 2027, while rules for high-risk AI embedded in regulated products are scheduled for 2 August 2028.[3]

Meer reports that critics warn compliance costs could hit startups hardest and make it harder for them to compete with larger technology companies.[1] Swedish Prime Minister Ulf Kristersson has called parts of the law confusing and burdensome, Meer reports.[1] Supporters argue that clear rules can build public trust and encourage acceptance of AI in sensitive areas such as healthcare or finance.[1] The European Commission says more than 180 organisations signed its Code of Practice on transparency of AI-generated content.[4] The European Parliamentary Research Service says delays in preparing required standards were among the reasons high-risk AI deadlines were postponed.[2]