An institutional framework for shared rules
The European Union was created when the Maastricht Treaty entered into force on November 1, 1993. Its member states work through EU institutions to make shared policy. In the AI Act’s legislative process, the European Commission proposed the legislation and the Parliament and Council acted as co-legislators. [1][2]
The AI Act has more than one legal milestone
The March 2024 Parliament vote was followed by Council approval in May, formal signature on June 13 and Official Journal publication on July 12. The regulation entered into force on August 1, 2024. These dates record different steps, rather than a single vote making all duties immediately applicable. [2][3]
The law addresses prohibited uses, requirements for high-risk systems, transparency and obligations for general-purpose models. The European Parliament’s October 2026 timeline also records the 2026 Digital Omnibus amendment. Current application dates therefore need to be read with the amendments, rather than assuming the original 2024 schedule remains unchanged. [2][4]
Application and enforcement are distributed
The Commission’s AI Office, national competent authorities and the European Data Protection Supervisor have different enforcement responsibilities. GPAI obligations, transparency rules and later high-risk requirements have separate stages. A shared EU regulation does not establish identical day-to-day enforcement in every member state. [5][4]