Meta signs the EU AI content transparency code (July 28, 2026)
Meta confirmed on July 28, 2026 that it signed the EU Code of Practice on AI content transparency. What the date means for teams making ad video.

Meta confirmed on July 28, 2026 that it is signing the EU AI Act Code of Practice on Transparency of AI-Generated Content, and says it has labeled AI content since February 2024. For a team making ad video, the signature changes nothing you can toggle, but it points the same way as Meta's existing AI info labels: plan for the label.
The dates
Three public pages give the sequence. All were read on 2026-10-03.
| Date | Event | Source |
|---|---|---|
| February 2024 | Meta says it began labeling AI content | Meta newsroom, July 2026 post |
| 1 June 2026 | Meta page updated: "About this ad" includes AI info labels | Meta newsroom, ads post |
| 10 June 2026 | Final Code of Practice on marking and labelling published | European Commission page |
| 28 July 2026 | Meta confirms it is signing the code | Meta newsroom |
| End of July 2026 | About 190 organisations signed the code | European Commission page |
| 2 August 2026 | Article 50 transparency obligations apply | European Commission page |
What Meta's ads page adds
Meta's ads transparency post, updated June 1, 2026, says "About this ad" includes AI info labels and that Meta automatically detects ads created or edited with third-party AI tools through industry-standard signals. It says labels sit in the three-dot menu, or next to "Sponsored" when photorealistic AI humans are present.
So for creative made with any AI video tool, including Sume, a label can attach whether or not you add one. The code signature is a statement of direction. It does not describe a new button or field for advertisers, and the pages read do not say so.
What a production team can do now
A few steps hold up regardless of how the rules develop:
- Keep one record per ad: which tool made it, the prompt, the date, and who approved it.
- Decide per ad whether a person in the frame is photorealistic, since that is the case Meta names for the label next to "Sponsored".
- Write the ad so it still works with a label showing. A claim that depends on the viewer believing the footage is filmed is the fragile one.
- Check your own country's disclosure rules for ads separately, since an EU platform label does not replace an advertiser's own duties.
Storing the record with the job id
Keep the record in your own storage, keyed by the Sume job id returned from the video API, so the approval trail survives into your reporting.
{
"job_id": "<job id>",
"ad_id": "spring-launch-03",
"tool": "sume/auto",
"prompt": "A product shot on a wooden table, soft daylight, slow pan",
"approved_by": "creative-lead",
"photoreal_person": "no"
}Sources
Related posts
More in Use cases
- Midjourney V8.2 edit stills, animated on Sume
Finish a still in Midjourney V8.2, host it at a public HTTPS URL, and use it as the first frame of a Sume video request. Midjourney itself has no video id here.
- H3 mixing camera, character and audio from three inputs
MiniMax H3 can mix camera movement from one video, a character from an image and audio from another. Sume's minimax-h3 takes image, video and audio references.
- Mixed-language explainer: two TTS jobs joined into one track
A voice that handles 23 or 90 languages still needs one language per request. Render each language as its own job, then join the takes with timeline audio.
- Explainer music that changes per scene: three Lyria briefs, one concat
Make each scene's music a separate Lyria 3.5 brief at least 12 BPM apart, then join them gaplessly with Timeline audio concat for $0.01.
Written by Sume