C2PA metadata

C2PA metadata is like a résumé attached to a photo or video file. It records inside the file whether the image was taken with a camera or generated by AI, and what tools were used to edit it afterward. For example, if a photo taken with a camera is later touched up with a photo-editing app, that process gets recorded as information you can check later.

It's a bit different from a watermark, which is invisibly embedded in a file. A watermark hides a signal inside the content, while C2PA metadata is more like a record card that logs the stages the content has gone through. That's why even when the on-screen watermark display can be turned off, C2PA metadata is designed to remain in place.

Since people can't read this information directly, browsers or apps read it for you. For example, if you right-click an image in Chrome and ask whether it was made by AI, Chrome reads this metadata and shows you details like the capture device or edit history.

The article notes that "the on-screen watermark display can be turned off, but the invisible SynthID itself and the C2PA metadata remain intact." A common point of confusion: C2PA metadata is not the same as the SynthID watermark—it's separate history information attached to the file. So even if you remove the visible watermark indicator, the C2PA metadata isn't erased and stays in place.

1. Prepare an image you want to check—whether it was taken with a camera or generated by AI.
2. In the Chrome browser, right-click on that image.
3. From the menu that appears, select "Is this image AI-generated?"
4. Chrome reads the C2PA metadata and watermark signals embedded in the file together, and shows you the answer, including details like the capture device or edit history.