Home > Blog > Google’s Gemini AI Sparks Controversy Over Watermark Removal
Google's Gemini AI Sparks Controversy Over Watermark Removal
Google’s latest AI model, Gemini 2.0 Flash, is making waves — and not entirely for the right reasons. Users across social media have discovered that the AI can be used to remove watermarks from images, including content from major stock media companies like Getty Images.
Gemini 2.0 Flash was recently updated to include an image generation and editing feature, giving users the ability to create and modify visual content with ease. While this capability is undeniably impressive, it appears to lack crucial safeguards. Reports indicate the model can generate images of celebrities, copyrighted characters, and, most controversially, erase watermarks from existing photos.
Fill The Gaps
What makes Gemini 2.0 Flash particularly noteworthy is its skill at filling in the gaps left behind when a watermark is removed. While other AI tools have dabbled in this area, Google’s model seems to excel — and it’s currently free to use.
It’s important to note that the image generation feature is still labeled as “experimental” and is only accessible through developer tools like Google AI Studio. Plus, the model isn’t flawless; it struggles with semi-transparent watermarks or those covering large portions of an image. Nevertheless, the ease with which watermarks can be erased is sparking concern among copyright holders.
Industry Response
Other AI companies, like Anthropic and OpenAI, have drawn a line in the sand when it comes to watermark removal. Anthropic’s Claude 3.7 Sonnet flat-out refuses to remove watermarks, deeming it “unethical and potentially illegal.” U.S. copyright law supports this stance, making unauthorized watermark removal illegal in most cases.
In response to the growing controversy, a Google spokesperson issued the following statement:
“Using Google’s generative AI tools to engage in copyright infringement is a violation of our terms of service. As with all experimental releases, we’re monitoring closely and listening for developer feedback.”
