Gemini Nano Banana Photos: How to Check if a Photo Was Made or Edited with AI
By Vikram Pagadala
· 5 min read
Google Gemini's image model — nicknamed Nano Banana — made AI photo edits a social-media trend in India, from retro saree portraits to 3D figurines. It also made it easy to fake photos that matter. Here's how to tell if a picture came out of Gemini.
Look for the ✦ sparkle watermark
Images made or edited in the Gemini app carry a small, semi-transparent four-pointed sparkle in the bottom-right corner. Popped! searches exactly that spot at every scale, so it finds the sparkle even after a photo was resized or forwarded on WhatsApp. Cropping the corner removes it, though — so its absence proves nothing.
Check the Content Credentials
Newer Google images also embed Content Credentials (C2PA) and an invisible SynthID watermark. Popped! reads C2PA data and shows when it declares generative AI. Screenshots and many apps strip this data, which is why the deep-learning check matters.
Run an AI image detector
When the watermark and metadata are gone, AI detectors trained on Nano Banana and other 2025–26 generators can still recognise the image. Upload it at popped.info — real photos come back Original, AI images get popped.