Advancing content provenance for a safer, more transparent AI ecosystem

OpenAI is strengthening its content provenance approach through three complementary layers: C2PA conformance (cryptographic metadata standards), SynthID watermarking via a Google DeepMind partnership for images generated in ChatGPT and the API, and a new public verification tool at openai.com/verify. The tool lets anyone check whether an image was generated by OpenAI by detecting both Content Credentials and SynthID signals. OpenAI acknowledges no single method is foolproof — metadata can be stripped and watermarks can fail — so the multi-layered approach is designed to be more resilient. The tool currently covers only OpenAI-generated content, with plans to expand cross-platform support.