OpenAI adds C2PA and Google’s SynthID to tag AI images
▼ Summary
– OpenAI announced two new measures to help the public identify images created by its AI models.
– The company is formally joining the C2PA open standard for content provenance and authenticity.
– OpenAI is partnering with Google to embed its invisible SynthID watermark across OpenAI’s image outputs.
OpenAI is rolling out two new tools to help people identify whether an image was generated by its AI systems. The company has officially joined the Coalition for Content Provenance and Authenticity (C2PA) open standard, and has also partnered with Google to embed its invisible SynthID watermark across all image outputs from OpenAI’s models.
These steps are part of a broader push to improve AI transparency and combat the spread of misleading or synthetic media. By adopting the C2PA standard, OpenAI will attach cryptographic metadata to images, indicating their origin and any edits made along the way. This makes it easier for platforms and users to verify authenticity.
The partnership with Google adds another layer of protection. SynthID is an imperceptible digital watermark that stays embedded in the image, even after cropping, resizing, or compression. Unlike visible watermarks, it’s designed to be machine-readable, allowing detection tools to flag AI-generated content without disrupting the visual experience.
Both measures aim to address growing concerns about deepfakes and misinformation as generative AI becomes more accessible. OpenAI’s move follows similar commitments from other major tech companies, signaling a shift toward industry-wide standards for AI content labeling.
The company says these features will be integrated into its image generation tools in the coming months, with broader support for detection and verification across third-party platforms expected to follow.
(Source: The Next Web)
