Google Ads AI labels: SynthID and C2PA cover Google-made assets
Google Ads embeds SynthID and C2PA in assets made inside its own tools. An MP4 made elsewhere and uploaded relies on your AI label setting instead.
Google's page says it embeds SynthID watermarks and C2PA markup in images and videos "generated within Google Ads tools". A video you generate elsewhere and upload is not described that way, so for that file you use the Ads AI label setting yourself. Nothing on the page says the setting adds a SynthID or C2PA mark to an upload.
Google's statements come from its AI content label help page, read 2026-09-30. Sume's side is from the API reference. This is not legal advice; Google's own note says the label setting does not guarantee compliance with any regulation.
What does Google mark, and what does it leave to you?
Two separate mechanisms appear on the page: a label you set, and machine-readable marking Google applies to its own tool output.
| Mechanism | Who applies it | What the page says |
|---|---|---|
| AI label setting | You | Designate assets as AI generated or edited; disclosure appears under "How this ad was made" in My Ad Center |
| Visible overlay | Google, from your designation | For campaigns targeting the European Union, India and New York |
| Labels Google adds | Where Google is legally required, or on signals from other platforms; "labels cannot be overwritten" | |
| SynthID and C2PA | Assets generated within Google Ads tools | |
| Your own label | You | Allowed in your own design tools; keep it off the edges and corners |
Does an uploaded MP4 get SynthID or C2PA from Google Ads?
The page ties the marking to generation inside Google Ads tools. It does not say Google adds SynthID or C2PA to a file generated somewhere else. What it does describe for your own assets is the AI label setting and the option to apply your own label.
What does a Sume video carry into Google Ads?
A Sume result is a Sume-hosted artifact: the reference says public results should use media.sume.com URLs, and the result shape holds an artifact id, url and type. The Sume docs I read do not describe SynthID or C2PA on that output, and Sume's output is not Google-tool output, so assume the file arrives without Google's marking. The per-model vendor pages are where a model's own marking, if any, is described; see does Sume add C2PA or a watermark to AI video.
What should I keep when I upload?
Keep the job id and artifact for each clip. GET /v1/jobs/:id/events returns public timeline events for debugging and recovery, which gives you a record of how the file was made. Then designate the asset as AI generated or edited in the Ads label setting, or add your own visible label, and check the rules for the regions you target. Other platforms differ; see the disclosure checklist by platform.
Sources
Related posts
More in Use cases
- Asset Studio video in 16:9 and 9:16: two Sume requests
Google's Asset Studio makes horizontal and vertical video in one workflow. With Sume, send one brief as two requests, 16:9 and 9:16, with model limits.
- Business Agent for YouTube ads: what Sume makes, what it doesn't
Google's Business Agent answers shopper questions inside YouTube ads. Sume does not build that agent; it makes the product video the ad runs.
- Google Merchant apparel images worn by people: reference-guided images
Google asks for apparel product images worn by people. gpt-image-2.5 on Sume takes up to 16 reference images, so a garment photo can guide the image.
- Google Merchant Center image size: 500x500 minimum, 1500x1500 advised
Google lists at least 500 x 500 pixels for product images and recommends 1500x1500 or above. Sume's gpt-image-2.5 can output 1536x1536, which meets the advice.
Written by Sume