Gemini can now tell if an image is AI-generated
What's the story
Google has added a new feature to the Gemini app that helps users identify AI-generated content. The update allows users to ask "Is this AI-generated?" to determine whether an image was created or modified using a Google AI tool. For now, the feature works only with images, but support for video and audio is expected soon.
Expansion strategy
Google's future plans for content verification
Google aims to bring this verification feature beyond the Gemini app and integrate it into its search engine as well. The company also plans to adopt industry-wide C2PA content credentials in the future, allowing users to verify content created by a range of AI tools and creative software, including OpenAI's Sora. At present, image verification is powered by SynthID, Google's own invisible AI watermarking technology.
Model update
Nano Banana Pro model and C2PA metadata
Google has also announced that images generated by its newly unveiled Nano Banana Pro model will contain C2PA metadata. This comes after TikTok's confirmation of using C2PA metadata for its own invisible watermarking of AI-generated content. The move is a step toward improving content verification across platforms and ensuring transparency in the use of AI tools for image creation or editing.