Google Gemini's “Banana Ai Saree Trend” transformed ordinary photography into a dramatic 90s Bollywood vintage style edit, creating quite a bit of talk on Instagram. Almost every girl appears to be trying out the AI-powered fashion trends, uploading virus edits with chiffon saris in the wind and warm golden lighting on social media. However, while Instagram users named Jhalakbhawani were trying out the trends, they found something “unsettled” and encouraged them to document their experience on Instagram.
What did users find “creepy”?
The woman shared that she tried to generate an image of herself and found something “creepy.” “On Instagram, I upload images to Gemini at one prompt, and Gemini converts them to Sally. Last night I tried this trend myself and found something very unsettling,” she writes.
She shared an image she uploaded to Gemini – herself in a green full sleeve suit – wrote a prompt on it. But the results Gemini produced shocked her.
Read Also | 10 Nano Banana AI Image Style Can be Created with Gemini, Google Reveals
She said: “This image is very appealing and I even posted it on Instagram. But I noticed something strange. In my left hand I have a mole that I actually have in real life. The original image I uploaded didn't have one.”
She further asked: “How did Gemini know that there is a mole in this part of my body? You can see a mole – it's very scary and creepy. I don't know yet how this happened, but I wanted to share this with you guys.
Her posts have sparked a lot of responses and have some safety concerns about trends that have set a few flags, but others have considered it “normal” and suggested that she made a video to get her opinion.
Also Read | 15 Vintage Retro Style AI Portraits Inspire You to Catch Viral Instagram Trends
Read | Gemini Ai Saree Photos: Best Prompts to Get Vintage Bollywood Vibes
How safe is the Gemini Nano Banana Tool?
Tech giants like Google and Openai provide tools to protect the content they use, but experts say safety ultimately depends on personal practices and intent to access images. For example, Google's Nano Banana image comes with an invisible digital watermark called Synthid, along with a metadata tag designed to “clearly identify as ai generation,” according to aistudio.google.com. It is not visible to the naked eye, but can be detected with special tools to verify the AI origin of images, SpielCreative.com reports.
Can watermarks really prevent misuse?
However, the detection tool has not yet been published. This means that most viewers cannot confirm the reliability of the images, Tatler Asia said. Critics also warn that watermarks can be easily forged or removed. “I don't think a watermark is enough,” said Honey Farid, professor of Berkeley, California, but Ben Colman, CEO of Real-World Defender, added that their real-world applications often “fail from onset.” Experts suggest combining other technologies with watermarks to better combat persuasive deepfakes.
