Article 50 draft transparency guidelines: what teams should track
The Commission opened a consultation on draft Article 50 transparency guidelines on 8 May 2026. The feedback window closed on 3 June 2026, and the underlying Article 50 rules still point to 2 August 2026 under the current law.
What the draft guidelines cover
The guidelines are meant to clarify scope and help providers and deployers of interactive and generative AI systems comply with Article 50. The consultation targets companies and organisations that develop or deploy AI systems interacting with individuals or generating synthetic content, including deepfakes.
What to do before final guidance
Teams should not wait for the final text to start practical disclosure work. The low-regret work is already clear: identify interactive AI touchpoints, define user-facing notices, decide where labels appear, and keep screenshot evidence of implementation.
- Map chatbots and assistants that interact directly with natural persons.
- Draft labels for deepfakes and AI-generated public-interest text.
- Plan machine-readable marking or detection support for generated content where relevant.
- Keep current-law timing visible: Article 50 still points to 2 August 2026.
FAQ
Is the consultation still open?
No. The Commission page lists the feedback window as 8 May to 3 June 2026. Teams should now monitor for final guidance.
Does this replace the Code of Practice?
No. The guidelines clarify Article 50 scope, while the Code of Practice is a voluntary support tool for marking and labelling AI-generated or manipulated content.
Turn this reading into an actionable report
Use the free scanner to map your likely role, detect likely obligations, and see which evidence is missing.