Google’s New AI Model Raises Concerns as Users Remove Watermarks from Images
People on social media have discovered a controversial use for Google’s latest AI model, Gemini 2.0 Flash—it can remove watermarks from images, including those from major stock photo providers like Getty Images.
Last week, Google expanded access to its Gemini 2.0 Flash model, allowing users to generate and edit images more seamlessly. This AI tool is undeniably powerful, but it also seems to lack some critical safeguards. Users have found that it not only erases watermarks but also fills in the missing areas with surprising accuracy. While other AI tools can do this too, Gemini 2.0 Flash appears to be exceptionally effective—and, crucially, free to use.
New skill unlocked: Gemini 2 Flash model is really awesome at removing watermarks in images! pic.twitter.com/6QIk0FlfCv
— Deedy (@deedydas) March 15, 2025
This discovery has sparked concerns among copyright holders. Watermarks exist to protect the rights of photographers, artists, and media companies, ensuring their work isn’t used without permission. However, with Gemini 2.0 Flash making it easy to erase them, many fear it could encourage unauthorized usage of copyrighted material.
Gemini 2.0 Flash, available in Google's AI studio, is amazing at editing images with simple text prompts.
— Tanay Jaipuria (@tanayj) March 16, 2025
It also can remove watermarks from images (and puts its own subtle watermark in instead 🤣) pic.twitter.com/ZnHTQJsT1Z
Google has labeled the image generation feature as “experimental” and “not for production use,” restricting its availability to developer-focused tools like AI Studio. Moreover, the AI isn’t perfect—it struggles with semi-transparent watermarks and those that cover large sections of an image. Even so, its ability to remove standard watermarks with such precision is alarming.
In contrast, other AI companies, like Anthropic and OpenAI, have stricter policies. Their models—Claude 3.7 Sonnet and GPT-4o—explicitly refuse to remove watermarks, calling the practice "unethical and potentially illegal." Under U.S. copyright law, removing a watermark without the owner’s permission is illegal in most cases.
The situation raises serious questions about the ethical and legal responsibilities of AI developers. Will Google implement stricter safeguards, or will this technology become another tool for digital piracy? As of now, Google has not responded to requests for comment.
