EU delays AI Act implementation as US debates its own comprehensive law
The EU postponed key deadlines of its AI Act by more than a year, citing unfinished technical standards, while US lawmakers are drafting a massive federal AI bill.
A new EU regulation, effective July 27, delays the enforcement timeline of the AI Act, moving high-risk AI compliance from August 2026 to December 2027 and extending deadlines for AI embedded in regulated products to August 2028. The postponement stems from incomplete harmonised technical standards and the lack of appointed national regulators, meaning the law’s definitions were not yet finalized. The European Commission acknowledges the administrative burden on businesses, especially small firms, and aims to cut such costs by 2029.
Meanwhile, US Congress is considering the Great American Artificial Intelligence Act of 2026, a 269-page proposal that would suspend state AI regulations for three years while adding federal obligations for frontier AI developers. Observers note that Europe’s struggle illustrates the difficulty of locking in detailed rules for a technology that evolves quickly, suggesting a narrower, conduct-based regulatory approach may be more effective.
Why it matters
The EU delay and US drafting reveal global uncertainty over how to regulate fast-moving AI technologies.
In this story
