The EU AI Act requires deployers of AI that generates deepfakes to disclose that the content was artificially generated or manipulated
Verified Confidence: high
Primary source found; it says the same thing, the scope matches, and it is reasonably current.
- Commonly cited as
- "EU AI Act: AI-generated content must be labeled"
- True origin
- European Union (AI Act, Art. 50) — Article 50: Transparency obligations for providers and deployers of certain AI systems
- Type
- Reference fact: A frequently needed number or rule taken directly from a primary source (e.g. Pew, Census, IAB/PwC, FTC), with its as-of date.
- Year
- 2024
- Data as of
- Regulation (EU) 2024/1689; Article 50 applies from Aug. 2, 2026
- Last checked
- What the source says
that generates or manipulates image, audio or video content constituting a deep fake
- Method & sample
- Regulation text. Providers must also mark synthetic outputs in a machine-readable format so they are detectable as artificially generated or manipulated.
- Drift between quote and source
- Scope: deepfakes and certain AI-generated text on public-interest matters; not every AI-assisted ad
- Exceptions exist for evidently artistic, satirical or fictional work
- Interest
- n/a (statute or regulation).
Safe rephrasing
Article 50 of the EU AI Act (applying from August 2, 2026) requires deployers to disclose AI-generated or manipulated deepfake image, audio or video content, and providers to mark synthetic output in machine-readable form.
Sources we opened
- EU AI Act Article 50 (primary)
- EU AI Act Article 113 (entry into application) (primary)
Machine-readable: /api/claims/eu-ai-act-deepfakes-must-be-disclosed.json · Last checked 2026-10-16