Ad
Skip to content

OpenAI supports C2PA standard for AI images and publishes classifier

OpenAI supports the C2PA standard for traceability in AI images and provides the first researchers with access to its own AI image classifier. The company joins the C2PA Steering Committee to help develop the digital content traceability standard. OpenAI adds C2PA metadata to DALL-E 3 images and plans to do the same for the Sora video model. In addition to supporting the C2PA standard, OpenAI is developing its own tamper-proof watermarks and AI-based recognition classifiers for detecting AI-generated images. The company has made its DALL-E 3 image classifier available to select research institutions. In internal tests, OpenAI's classifier correctly identified 98 percent of the images generated by DALL-E 3, while mislabeling less than 0.5 percent of non-AI-generated images. However, the classifier identified only 5 to 10 percent of images generated by other AI models as AI-generated. According to OpenAI, collective industry approaches are needed to effectively address the issues associated with AI-generated images.

AI News Without the Hype – Curated by Humans

Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section.

Read on for the full picture.
Subscribe for hype-free coverage.

  • Full access to every article on THE DECODER
  • No ads
  • Join the comments and community discussions
  • A weekly AI news recap via mail
  • 6x/year: "AI Radar" — deep dives on the AI topics that matter most
  • Daily AI news, always up to date
  • Our full ten-year archive
  • Covered by a team with 10+ years in AI
Subscribe to The Decoder