Advancing content provenance for a safer, more transparent AI ecosystem
OpenAI Blog · 2026-05-19
OpenAI announces a multi-layered content provenance system combining C2PA conformance, Google DeepMind's SynthID invisible watermarking, and a new public verification tool to help users identify AI-generated images.
Appears in
Extraction
Topics: content-provenanceai-watermarkingc2padigital-trustopenai
Claims
- OpenAI has achieved C2PA Conforming Generator Product status, allowing platforms to reliably read and preserve provenance metadata attached to OpenAI-generated content.
- OpenAI is integrating Google DeepMind's SynthID invisible watermarking into images generated via ChatGPT, Codex, and the OpenAI API.
- A new public verification tool allows anyone to check whether an uploaded image contains OpenAI provenance signals, including Content Credentials and SynthID watermarks.
- C2PA metadata alone is insufficient because it can be stripped or lost through file transformations such as screenshots, resizing, and format changes.
- No single provenance technique is sufficient; combining open standards, durable watermarking, and public verification tools creates a more resilient ecosystem.
Key quotes
Provenance only works if it survives beyond the first platform where content is created, and conformance makes that possible.
These two systems reinforce each other. C2PA helps content carry detailed context; SynthID helps preserve a signal when metadata does not survive.
No single provenance technique is enough on its own. We believe a strong approach combines shared standards, durable watermarking signals, and public verification.