You can now use the Gemini app to spot AI-generated images

As AI images get more realistic—we’re looking at you, Nano Banana Pro—it is becoming increasingly difficult to trust your eyes. Especially when it comes to news, politics, or even product reviews, knowing if an image is synthetic is a big deal.

Today, Google announced a new feature that puts a “truth detector” directly inside the Gemini app. According to the official blog post, you can now upload an image to the Gemini app and simply ask, “Was this created with Google AI?” or “Is this AI-generated?”

How it works

Gemini will analyze the image specifically looking for SynthID, Google’s invisible digital watermarking technology. If it finds the watermark, it will tell you that the image was generated or edited by Google AI.

There is a catch, though. Right now, this only works for images created with Google’s own tools. Because it relies on the proprietary SynthID watermark, Gemini can’t yet tell you if an image was made by Midjourney, OpenAI, or other competitors.

The “Nutrition Label” for AI is coming

Google also announced that starting this week, images generated by the new Nano Banana Pro model will include C2PA metadata.

If you aren’t a tech nerd, think of C2PA as a “digital nutrition label.” It’s an open industry standard (supported by Adobe, Microsoft, and others) that attaches a permanent record to a file showing where it came from and if it was edited by AI.

While the current “Ask Gemini” feature relies on Google’s own watermark, Google confirmed they plan to support verifying these universal C2PA credentials in the future, which would eventually let you check the origin of content from outside Google’s ecosystem, too. For now, if you suspect a photo is a Google AI creation, just drop it into Gemini and ask.