C2PA writes a manifest into the asset: which tool made it, what was changed, and by whom, each step cryptographically signed. Strip it and the absence is itself information. It is the closest thing the industry has to a chain of custody for a picture.
For a brand this is becoming a procurement question rather than an ethics question. Platforms, broadcasters and an increasing number of client legal teams now ask what provenance an asset carries before it runs. Answering "none" is a slower answer than it used to be.
We treat credentials as a deliverable, not an afterthought. If a client needs them attached, that goes in the brief, because retrofitting provenance onto a finished cut is not possible.
Is C2PA the same as a watermark?
No. A watermark marks the pixels. C2PA signs the file’s history. A watermark survives a screenshot badly; a C2PA manifest does not survive a screenshot at all, which is why serious workflows use both.
Do I legally have to disclose AI-generated content?
It depends on where it runs and what it claims. Several platforms require a synthetic-media label, the EU AI Act sets transparency duties for certain deepfake content, and advertising codes apply regardless. We work to the stricter of platform policy and client legal.