YouTube's Made with AI label and How this content was made

YouTube's label for meaningfully altered or synthetic content reads Made with AI, and it appears in a How this content was made section of the expanded description. YouTube also carries forward disclosures from tools whose files hold secure C2PA data (version 2.1 or higher). Sume's docs do not describe C2PA, so for a Sume file the dependable route is to answer the AI use question yourself in YouTube Studio.

What does the page say? (read 2026-10-09)

What does a Sume file carry?

The video tool docs cover the probe, stills and transcript; they say nothing about embedded provenance data. So do not count on YouTube reading a Made with AI or a camera mark from a Sume export. Probe your own delivered file with a C2PA reader if that matters for your workflow, as in what a trim or re-encode does.

What should you do instead?

• At upload, answer the AI use question in Studio under Attributes: Yes or No.
• Answer Yes for realistic people, events, scenes, or music that is a main focus, per the disclosure page.
• Skip it for the exempt cases the page lists, such as beauty filters, colour and lighting, caption creation and audio repair.
• Keep the project log of which clips came from a model and which from a camera.

What happens if you skip it?

The disclosure page says creators who consistently choose not to disclose may get a label applied by YouTube or penalties including removal of content or suspension from the YouTube Partner Program. See what YouTube can do for the full list.

If you use YouTube's own generative tools for part of a video, expect the label to be applied for you. For everything else, the answer in Studio is yours to give, and the safest habit is to answer from your project log rather than from memory on upload day.