Regulation
64 analyses
The European Commission's AI Office has assumed formal powers to oversee and fine general-purpose AI (GPAI) providers as of this Sunday. Meta is the only major Western company to refuse the Code of Good Practices.
The 12-month adaptation period for general-purpose AI providers ends on August 2, 2026. Starting this date, the EU AI Office can impose fines of up to 3% of global revenue on models such as those from OpenAI, Meta, and Google DeepMind.
Regulation 2026/1744 maintains the enforceability of Article 50 on August 2, postpones high-risk obligations to 2027, and expands the AI Office. Fines of up to 15 million euros or 3% of global turnover.
Published as Regulation 2026/1744 and effective since July 27, the AI Omnibus postpones high-risk obligations by 16 months but maintains intact the transparency rules that take effect on August 2.
Binding DMA specification published on July 16 forces Google to share search data with competitors starting January 2027 and unlock 11 Android features for rival AI assistants.
The CAC has authorized Apple Intelligence in China to operate on Alibaba and Baidu models, a precedent that reshapes the global competitive landscape of AI in devices.
Executive Order No. 62 freezes 39 pending applications, requires a new environmental assessment, and tests whether the AI network will reach regulatory limits before physical limits.
With the package approved by the European Council on June 29, the horizon for high-risk AI extends to December 2027, but the deadline of August 2 for transparency in chatbots remains unchanged.
The General Court of the EU rejected Apple's appeals against the designation of gatekeeper and upheld the Commission's path to regulate Google, Meta, Amazon, Microsoft, and ByteDance.
CNBC report reveals an investigation after companies admitted to using Qwen and Kimi in their tech stack, with the share of Chinese models in OpenRouter traffic fluctuating between 30% and 46% weekly since February.
The UN Secretary-General called for international safeguards for frontier models, a ban on lethal autonomous weapons, and a Global AI Fund in a speech in Geneva.
Meeting at Palexpo, with 193 member states co-chaired by El Salvador and Estonia, aims to establish a multilateral floor as the US, European Union, and China pursue divergent paths.