Google’s AI Model Can Remove Watermarks, Raising Copyright Concerns

• Mar 17, 2025

Google’s new AI model, Gemini 2.0 Flash, has sparked controversy after social media users discovered it can remove watermarks from images with remarkable accuracy.

AI Model Raises Copyright Issues

The AI tool, recently introduced by Google, has been found to not only erase watermarks but also intelligently fill in the gaps left behind, making images appear untouched. While AI-powered watermark removal is not new, Gemini 2.0 Flash appears to outperform other models in this capability, according to reports by TechCrunch.

This development has raised concerns among photographers and artists, who rely on watermarks to protect their work from unauthorized use. Watermarks serve as proof of ownership and are typically removed only when someone legally purchases an image. The ability of AI to bypass this security measure threatens the livelihood of content creators who depend on selling their images.

Google’s AI Model Is Still in Testing

Currently, Google has only made Gemini 2.0 Flash available through its developer tools, and the company has labeled the image generation feature as "experimental" and "not for production use." However, social media users on platforms like X (formerly Twitter) and Reddit have shared examples of the AI removing watermarks, raising fears about potential misuse.

Although the AI struggles with some types of watermarks, particularly semi-transparent ones, it is still highly effective. This has led to calls for stronger protections to prevent AI tools from being used for copyright infringement.

Legal and Ethical Concerns

Removing a watermark without permission is considered copyright infringement under U.S. law and is illegal in most cases. If AI makes it easy to bypass these protections, it could create widespread legal issues and financial losses for artists.

Even if Google implements safeguards to limit this functionality, other AI tools are likely to emerge with similar capabilities. The concern is not just about Google’s model but about the broader implications of AI’s ability to manipulate copyrighted content.

As AI technology advances, there is growing pressure on tech companies to establish stricter controls to protect intellectual property rights and ensure that artists and photographers are compensated for their work.