OpenAI implements C2PA standards and SynthID watermarking
openai adopted c2pa standards for content credentials and integrated google deepmind's synthid watermarking for images generated by chatgpt and the api.
openai strengthened content provenance by achieving c2pa conformance for dall-e 3, sora, and imagegen, allowing metadata to travel securely across platforms. they partnered with google deepmind to implement synthid, an invisible watermarking layer for images, ensuring durability even when metadata is stripped by screenshots or format changes. this dual-layer system protects against manipulation and provides verifiable context for journalists and platforms. additionally, openai released an early developer preview of a media detection tool to verify content origins.