EU AI Act Hits Key August Enforcement Milestone, With High‑Risk Rules Pushed to 2027

The European Union’s landmark AI Act has reached a turning point: as of 2 August, key enforcement powers and transparency duties are now live. The European Commission and national market‑surveillance authorities can impose substantial fines—up to 7% of global turnover—for prohibited AI practices, and general‑purpose AI providers face binding obligations around documentation and transparency. Companies must also ensure users are clearly informed when they are interacting with an AI system, tightening rules around chatbots and other automated interfaces.

However, the most detailed requirements for so‑called “high‑risk” AI systems—covering areas like employment, education, biometrics and critical infrastructure—have been pushed back. A recently adopted amending regulation moves the compliance deadline for these Annex III use cases from August 2026 to 2 December 2027, citing unfinished technical standards and implementation work. For vendors, that delay offers extra time to align quality‑management, testing and logging processes. For civil‑rights advocates, it postpones stronger protections against opaque automated decision‑making in sensitive domains. For Japanese firms operating in Europe, the shift means AI governance can no longer be treated as a future concern: enforcement has begun, and non‑compliance now carries immediate legal and financial consequences.