Google has launched a new feature in Gemini 3, enabling users to identify whether an image is AI-generated. This tool leverages Google’s SynthID, a digital watermarking technology that embeds signals into AI content. To use the feature, simply upload an image and ask, “Is this image AI-generated?” Gemini’s verification is based on the presence of SynthID. Since its introduction in 2023, over 20 billion pieces of AI content have been watermarked. Google also plans to integrate the Coalition for Content Provenance and Authority (C2PA) standards, enhancing provenance tracking for AI-generated images beyond Google’s ecosystem.
Testing Gemini revealed promising results: it accurately identified a non-AI-generated photo and recognized AI-generated images—even when not utilizing SynthID—by detecting common traits. Users are encouraged to add C2PA signatures to their images using tools like Lightroom or Photoshop, enhancing content authenticity and transparency. This new feature aids users in navigating the landscape of AI-generated content.
Source link
