Google’s New AI Removes Watermarks

If you create content online—whether you’re a food blogger, photographer, or small business owner—you’ve probably relied on watermarks to protect your images from being stolen.

Google released a new AI model, Gemini 2.0 Flash Thinking Experimental (a ChatGPT and Grok competitor), and it has an unsettling ability: it can remove even complex watermarks from images. That’s bad news if you rely on watermarks to protect your work.

I had to see it for myself, so I tested it on an image I had permission to use and added a very simple watermark. The result? The watermark was gone in seconds:

Watermarks have never been bulletproof. Sure, someone with Photoshop skills could remove them, but it took time and effort. Now? AI can erase them in seconds, with no skill required.

Stock photo platforms like iStockPhoto, Getty, Shutterstock, and Unsplash+ have long used intricate watermark designs to discourage unauthorized use. But with this AI tool, even those layers of protection are at risk. As more people become aware of this capability, watermarking images may soon be meaningless. I want you to know now so you won’t be caught off guard.

Want to see for yourself? You can test it with your own images here:
Google AI Studio.

• Select Gemini 2.0 Flash (Image Generation) Experimental as the model (right-hand side).
• Choose Images and text as the output format.

I’d love to hear your thoughts on this. I hope you find this helpful (if sobering).
— Rachel