Since August 2, 2026, the EU AI Act requires providers of generative AI to mark what it makes as AI-generated in a machine-readable way, and deepfakes to be disclosed by whoever publishes them. Generative systems that were already on the market have until December 2, 2026 to add machine-readable marking.
Any image you govern can carry a visible and embedded disclosure label per jurisdiction (EU, California, New York).
Signed C2PA content credentials and a verifiable fingerprint on any image, video, or audio file you govern — across every model you use, not just one lab's.
A tamper-evident, hash-chained trail of what was governed, how, and what was disclosed — the record legal will ask for.
Each lab signs its own outputs. But a brand generating with five different models has no single way to sign it, prove where it came from, and keep one audit trail. That's the layer Varosity is building — one API across every model, where any asset can be signed and logged in a single call (images also get a disclosure label). Signing currently uses a pilot certificate.
Powered by the Varosity Trust Layer (pilot) · C2PA content credentials · governance for AI media across every model. This page is informational and not legal advice.