Privacy concerns raised over Google's Nano Banana gen AI photo generator

After ChatGPT's Ghibli-inspired AI photos craze in April earlier this year, Google has kicked off another similar trend on social media platforms in India.

In late August, Google updated its Gemini AI app with the advanced Gemini 2.5 Flash Image large language model, code-named Nano Banana.

With the latest Gemini AI app version, users can play with photos using the Reimagine feature. It can turn any mundane-looking photo into a fun, engaging piece of art.

The Gemini app can help users switch their outfit and even the background scene in a few seconds. Users can upload their own image or the pet and ask Gemini to reimagine the photo.

People are using to Gemini AI app to generate 3D figurines in a toy box in mint condition, and also come up with creative prompts to turn a photo into anime, manga, or mecha figures.

It has gained a lot of traction among the tech-savvy netizens. However, a user on Instagram has raised privacy concerns over the Gemini app, reported Livemint.

Google Nano Banana AI tool feature.

The user who goes by the name @jhalakbhawani tried the latest 'nano banana' trend ', Banana AI Saree'. On the app, she uploaded a photo of herself and asked it to generate her image with a retro saree with a sleeveless blouse on her.

Initially, she was thrilled to see the result. But, upon closer inspection, she noticed that the app was able to generate a birthmark on her shoulder. In the uploaded photo, she had a green full-sleeve suit, fully covering her shoulders. She was startled by how the Gemini AI app was able to show a birthmark at the exact point near the shoulder.

There is a possibility that the photo might just be a coincidental glitch, but privacy experts have long been warning people not to fall for such social media trends.

Big tech companies need data to train their generative AI models. Every form of information, be it in audio, video, images or just text, they are fed into Large Language Models, which are used to run AI applications.

Big technology companies are a little more careful, as legal ramifications will not only carry hefty fines, but also risk losing their global brand image.

Google spokesperson has clarified that the company has not trained its Nano Banana image model from any of its affiliated services such Google Photos and others. It is just a co-incidence that the image has shown an birthmark which was not visible in the uploaded photo.

“The Nano Banana image model that was recently launched was not trained with user data from Google Photos, Google Workspace, or Google Cloud services. Any non-visible attribute that shows up in a generated image that wasn’t visible in an input image in the same Gemini conversation is a coincidence.”

Google spokesperson

However, there are several low-profile firms and even cybercrime mercenaries who will come up with similar applications just to get personal information of people and target their devices to steal financial details or personal information for ransom. Or sometimes sell them to the highest bidder on the dark net.

So, people are advised not to follow social media trends just because a friend or family member or colleague participated in it. Be wary of such gen AI app promotions on social media platforms.

Get the latest news on new launches, gadget reviews, apps, cybersecurity, and more on personal technology only on DH Tech.