On August 2, the European Commission’s AI Office began enforcing Article 50 of the EU AI Act, and with it a transparency regime that reaches far beyond frontier labs. Chatbots must disclose they’re machines. Deepfakes need labels. Synthetic content requires machine-readable marks. AI-generated text on matters of public interest must be identified. Fines run up to €15 million or 3% of worldwide annual turnover, whichever is higher, jointly enforced by the AI Office, national market surveillance authorities, and the European Data Protection Supervisor.
Cooley’s read of the text is that Article 50 applies immediately to all in-scope systems regardless of when they were placed on the market. Providers of pre-existing generative systems get until December 2, 2026 to meet the machine-readable marking requirement, and that’s the extent of the runway.
The scope is what matters. Under Article 50, the deployer category sweeps in the no-code middle layer, platforms like Glean, Dust, and LemonLime that embed Claude or ChatGPT into ordinary business workflows. These aren’t model builders. They’re the substrate through which most European SMBs will actually touch generative AI, and now they carry disclosure obligations scaled to global revenue.
Brussels sequenced this deliberately. The heavier Annex III high-risk regime slid to December 2, 2027, and Annex I embedded products to August 2, 2028, under the AI Omnibus political agreement reached May 7. Guidelines on Article 50 scope landed July 20. The Commission’s voluntary Code of Practice on Transparency of AI-Generated Content drew more than 180 signatories at launch; per Wilson Sonsini, signatories receive a presumption of conformity and softer enforcement, while non-signatories must document alternative measures. A push to expand third-party AI evaluation capacity is expected operational by 2027.
The high-risk fight was postponed. The disclosure fight starts now, and it starts at the deployment layer where the regulation is most legible to ordinary businesses.
Sources
- Commission starts enforcing AI Act rules and new transparency requirements on 2 August, European Commission
- Safer and more transparent AI, European Commission
- AI Act, Shaping Europe’s digital future
- EU AI Act: Transparency Obligations Take Effect 2 August 2026, Cooley
- EU AI Act Enforcement Phase Begins, Wilson Sonsini
- LemonLime
Sources
- Commission starts enforcing AI Act rules and new transparency requirements on 2 August — European Commission
- Safer and more transparent AI — European Commission
- AI Act — Shaping Europe's digital future
- EU AI Act: Transparency Obligations Take Effect 2 August 2026 — Cooley
- EU AI Act Enforcement Phase Begins — Wilson Sonsini
- LemonLime