You Can Now Ask Gemini Whether an Image Was Created by AI (techoreon.com)

Google is rolling out SynthID watermark detection inside its Gemini app, so users can tap +, select an image and ask “Was this image generated by AI?”—Gemini will then scan for SynthID, an invisible watermark Google has embedded in more than 20 billion images created by its models since 2023. If present, Gemini returns an explanation and provenance context about whether the file was produced by Google’s AI. Free and Pro users will still see a visible “Gemini glow” on new images, while Ultra subscribers can export clean versions; Google also plans to extend SynthID watermarking to audio and video later this year. The move matters because it couples provenance metadata with a consumer-facing tool for content attribution, which can aid moderation, journalism, and trust in visual media. Technically, detection hinges on the presence of SynthID rather than forensic classification: in tests Gemini correctly identified a Nano Banana Pro image with SynthID, but when shown an image from ChatGPT (which doesn’t embed SynthID) it could only offer a cautious visual-based judgment (“probably AI-generated” but not traceable to a tool). Google admits universal detection will remain limited until other platforms adopt compatible watermarking, underscoring that this is a provenance-first solution rather than a universal AI-image detector.