Overview
- Instagram’s CEO proposes marking real photos and videos at creation or upload rather than relying on catching every AI-generated fake.
- The company says it is examining a cryptographic provenance approach that could assign a secure identifier to original media.
- Possible implementations under review include C2PA-style chains of custody and collaboration with camera and device makers to sign content.
- Instagram has not committed to a rollout or timeline, and details on scope and standards remain undecided.
- Mosseri argues labels must be paired with richer context about media and the accounts sharing it as AI-made posts surge in volume and realism.