Google Gemini's latest "Banana AI saree trend" is taking Instagram by storm. The tool transforms everyday photos into vintage 90s Bollywood-style edits, with chiffon sarees, dramatic poses, and golden-hour lighting. While the trend has gone viral, one user's unusual experience has raised serious safety concerns.
Instagram User's 'Creepy' Experience
An Instagram user named @Jhalakbhawani shared that she joined the trend by uploading a picture of herself in a green full-sleeve suit. She wrote a prompt asking Gemini to generate a saree look. At first, the result looked beautiful, and she even posted it online.
But then she noticed something unsettling: the Al-generated image showed a mole on her left hand-a mole she actually has in real life, but which wasn't visible in the original uploaded photo.
"How did Gemini know that I have a mole on this part of my body? The original photo didn't show it," she wrote. "It's very scary and creepy. I'm still not sure how this happened, but I wanted to warn others—please be careful with what you upload on AI platforms.
Social Media Reactions
Her post quickly spread online, with some users echoing her fears and urging caution, while others dismissed it as a coincidence or a way to get more views. Still, the story has sparked a broader discussion about privacy and AI safety.
Experts on AI Risks
Google says it uses SynthID, an invisible watermarking system, along with metadata tags to mark AI- generated content. However, experts point out these safeguards are limited. SynthID detection tools are not yet publicly available, making it difficult for everyday users to verify Al content.
Cybersecurity experts advise users to:
• Avoid uploading sensitive or personal images to Al platforms.
• Remove metadata, such as location tags, before sharing pictures.
• Keep a copy of original images to cross-check for unauthorised modifications.