Story
The EU AI Act's full enforcement for General-Purpose AI (GPAI) model providers activates on August 2, 2026, granting the European AI Office binding enforcement powers under Articles 88–91. From this date, GPAI providers face fines up to €15 million or 3% of global annual turnover under Article 101. GPAI models released before August 2025 must be fully compliant by this deadline, marking the end of the grace period. The European Commission introduced a coordinated framework in July 2025—including a GPAI Code of Practice (transparency & copyright obligations under Article 53, systemic risk safety under Article 55), a Supplementary Guidance document clarifying 'provider' scope and systemic risk thresholds, and a mandatory Training Data Disclosure Template—to establish pre-enforcement governance expectations. While the Code of Practice is formally voluntary, it acts as a quasi-regulatory compliance bridge, and adherence creates a presumption of compliance with the Act's GPAI provisions. The European AI Office, AI Board, Scientific Panel, and Advisory Forum are collectively responsible for enforcement, with major model providers having participated in the Code's development. Medialaws.eu reported the core details. Reg Intel noted additional context. ISAR Global added corroborating details.