European AI Office

2 Aug 2026 — Article 50 transparency obligations apply — and enforcement begins The date held. Providers and deployers must disclose: direct human–AI interaction; AI-generated or manipulated content (machine-readable marking); emotion recognition and biometric categorisation; and deepfakes and AI-generated text published on matters of public interest. These duties are not limited to high-risk systems and survived the Omnibus substantially intact — they are now the AI Act's operative compliance frontier, backed by penalties of up to €15m or 3% of worldwide turnover. On the same date the AI Office and national market surveillance authorities began enforcing the Act, and the Commission published a first list of more than 180 organisations that have signed the Code of Practice on Transparency of AI-Generated Content. Transitional relief is narrow: content published before 2 August 2026 need not be retrospectively labelled, and only the marking-and-detection duty for generative systems already on the market is deferred, to 2 December 2026. European Commission — Commission starts enforcing AI Act rules and new transparency requirements on 2 August (31 Jul 2026) <https://digital-strategy.ec.europa.eu/en/news/commission-starts-enforcing-ai-act-rules-and-new-transparency-requirements-2-august> European Commission press release IP/26/1714 — AI Act enforcement and transparency rules <https://ec.europa.eu/commission/presscorner/detail/en/ip_26_1714>

1 Sep 2026 — AI Office sends first enforcement requests for information to 30+ general-purpose AI providers A month after its GPAI enforcement powers applied, the Commission confirmed it had sent its first requests for information under the AI Act to more than 30 providers of general-purpose AI models — among them OpenAI, Anthropic and Google — in two streams: safety and security (including independent external evaluation and post-deployment monitoring for the most capable models) and copyright and transparency, including training-data summaries. RFIs are an investigative step, not a finding; the fining power behind them reaches €15m or 3% of worldwide turnover. This is the first observable use of the AI Office's direct supervisory powers. Agence Europe — European Commission sends first requests for information to more than 30 AI providers (Sep 2026) <https://agenceurope.eu/en/bulletin/article/13929/31/european-commission-sends-first-requests-for-information-to-more-than-30-ai-providers> Wilson Sonsini — EU AI Act enforcement phase begins <https://www.wsgr.com/en/insights/eu-ai-act-enforcement-phase-begins.html> CDT Europe — AI Bulletin: September 2026 (24 Sep 2026) <https://cdt.org/insights/cdt-europes-ai-bulletin-september-2026/>

1 Sep 2026 — AI Office sends first enforcement requests for information to 30+ general-purpose AI providers A month after its GPAI enforcement powers applied, the Commission confirmed it had sent its first requests for information under the AI Act to more than 30 providers of general-purpose AI models — among them OpenAI, Anthropic and Google — in two streams: safety and security (including independent external evaluation and post-deployment monitoring for the most capable models) and copyright and transparency, including training-data summaries. RFIs are an investigative step, not a finding; the fining power behind them reaches €15m or 3% of worldwide turnover. This is the first observable use of the AI Office's direct supervisory powers. Agence Europe — European Commission sends first requests for information to more than 30 AI providers (Sep 2026) <https://agenceurope.eu/en/bulletin/article/13929/31/european-commission-sends-first-requests-for-information-to-more-than-30-ai-providers> Wilson Sonsini — EU AI Act enforcement phase begins <https://www.wsgr.com/en/insights/eu-ai-act-enforcement-phase-begins.html> CDT Europe — AI Bulletin: September 2026 (24 Sep 2026) <https://cdt.org/insights/cdt-europes-ai-bulletin-september-2026/>

1 Sep 2026 — AI Office sends first enforcement requests for information to 30+ general-purpose AI providers A month after its GPAI enforcement powers applied, the Commission confirmed it had sent its first requests for information under the AI Act to more than 30 providers of general-purpose AI models — among them OpenAI, Anthropic and Google — in two streams: safety and security (including independent external evaluation and post-deployment monitoring for the most capable models) and copyright and transparency, including training-data summaries. RFIs are an investigative step, not a finding; the fining power behind them reaches €15m or 3% of worldwide turnover. This is the first observable use of the AI Office's direct supervisory powers. Agence Europe — European Commission sends first requests for information to more than 30 AI providers (Sep 2026) <https://agenceurope.eu/en/bulletin/article/13929/31/european-commission-sends-first-requests-for-information-to-more-than-30-ai-providers> Wilson Sonsini — EU AI Act enforcement phase begins <https://www.wsgr.com/en/insights/eu-ai-act-enforcement-phase-begins.html> CDT Europe — AI Bulletin: September 2026 (24 Sep 2026) <https://cdt.org/insights/cdt-europes-ai-bulletin-september-2026/>