OpenAI adds SynthID and C2PA signals to generated images

OpenAI's May provenance update is a practical synthetic-media infrastructure story. The company says it has become a C2PA conforming generator product, is incorporating Google DeepMind's SynthID watermark into images generated through ChatGPT, Codex, and the OpenAI API, and is previewing a public verification tool. C2PA can provide structured metadata about how a file was made or edited, while SynthID is designed to survive some transformations that strip metadata. The important limitation is that provenance is a signal, not proof of truth: files can lack credentials for innocent reasons, and determined adversaries can still use tools outside participating ecosystems. Still, OpenAI adopting a Google watermarking standard is a meaningful step toward cross-platform media provenance.

Key details: OpenAI, May 19, 2026, C2PA conforming generator, Google DeepMind SynthID, ChatGPT images, Codex images, OpenAI API, public verification preview.

Continue swiping for more AI Brief stories.