Story
The EU AI Act enters its next enforcement phase on August 2, 2026, when mandatory obligations for general-purpose AI (GPAI) model providers become effective under Regulation 2024/1689. Major model providers including OpenAI, Google, Anthropic, and Meta must implement mandatory disclosures about model capabilities, training data summaries, and technical specifications. The European AI Office will oversee compliance, with penalties for non-compliance reaching up to 3% of global annual turnover. A one-year grace period from August 2025 means no penalties will be imposed until August 2, 2026, giving providers time to align with the Code of Practice. TechRepublic reported the core details. European Commission Digital Strategy noted additional context. Securiti.ai added corroborating details.