Synthetic media governance is no longer just a trust-and-safety preference. Article 50 creates specific transparency duties for providers and deployers of systems that generate or manipulate content. The core operational problem is proving that the right marking or disclosure was attached to the right content, by the right actor, at the right time.
Article 50 establishes transparency duties for AI-generated and manipulated content, including machine-readable marking by certain providers and disclosure duties for certain deepfake and public-interest content. The Regulation also contains tailored treatment for artistic and similar works and an editorial-control exception for specified public-interest text.
Start with the actual system, role, use case and evidence boundary. The same regulation can produce different obligations for different actors and systems.
Article 50 distinguishes provider-side marking duties from deployer-side disclosure duties. Start by identifying whether the system generates or manipulates synthetic audio, image, video or text content and who is responsible for putting that content into use.
Providers of AI systems that generate synthetic audio, image, video or text content must ensure outputs are marked in a machine-readable format and detectable as artificially generated or manipulated, subject to the Regulation’s technical-feasibility and proportionality conditions.
Deployers of systems that generate or manipulate deepfake image, audio or video content must disclose that the content was artificially generated or manipulated. Article 50 also addresses AI-generated or manipulated text published to inform the public on matters of public interest.
Preserve the system version, content type, generation or manipulation event, marking method, user-facing disclosure, editorial process where relevant, publication context, exceptions relied on and the evidence showing the disclosure matched the actual content.
The goal is not merely to reach an answer. It is to preserve what facts, evidence, scope and limitations supported that answer at that time.
Separate provider duties from deployer duties and distinguish synthetic text, image, audio and video. Determine whether the content qualifies as a deepfake, public-interest text, artistic work or another category.
Bind machine-readable marking, visible or otherwise appropriate disclosure, and any editorial-control exception to the exact content and publication context rather than relying on a generic website disclaimer.
Record the model and system version, content provenance, marking technology, disclosure text or signal, publication state, responsible operator and any human editorial review so the transparency action can be reconstructed later.
A new generator, edited media, changed watermark or metadata process, new channel, removed disclosure, new editorial workflow or republishing event can make the prior transparency evidence stale.
Choose the smallest operating tier that fits the portfolio today. Upgrade when system count, team size or governance scope actually requires it.
Keep a living system-level evidence record with obligations, gaps, versions and revalidation state.
START EVIDENCE PASSPORT →Coordinate evidence, owners, documentation, incidents and team compliance work in one governed workspace.
START COMPLIANCE WORKSPACE →Operate broader high-risk, GPAI, FRIA, post-market and material-change governance across a growing portfolio.
START GOVERNANCE PRO →Run institutional governance with expanded users, authority workflows, examiner rooms and portfolio reporting.
START INSTITUTION →Use the free classifier to establish the system, intended purpose, possible actor role, EU exposure and unresolved facts. When continuing evidence infrastructure is needed, paid access begins at $19 per month. Independent human readiness review remains a separate service.
Provider-side duties include ensuring certain synthetic audio, image, video and text outputs are marked in a machine-readable format and detectable as artificially generated or manipulated. Deployer-side duties include disclosure for certain deepfake content and certain AI-generated or manipulated public-interest text.
Yes. Deployers of AI systems that generate or manipulate image, audio or video content constituting a deepfake must disclose that the content has been artificially generated or manipulated, subject to the Regulation’s specific treatment for artistic, creative, satirical, fictional and analogous works.
No. Article 50 specifically addresses AI-generated or manipulated text published to inform the public on matters of public interest. The Regulation also includes an exception where the AI-generated content has undergone human review or editorial control and a natural or legal person holds editorial responsibility for publication.
A generic disclosure may not prove that the relevant person encountered the required transparency at the right content and context. The better governance record ties the disclosure or marking to the actual content, publication event, system version and operator role.
Article 50 transparency obligations apply from 2 August 2026 under the AI Act’s application timeline.
No. TA-14 can preserve the content classification, operator role, marking and disclosure evidence, exceptions, gaps, publication state and revalidation history. It does not provide legal advice, certification, conformity assessment or regulatory approval.
TA-14 Exchange Activity
Live cumulative activity recorded across the public Exchange surface.
···
Visitors
Recorded public visitors
···
Page Views
Recorded Exchange views