Google’s latest AI model, Gemini 2.0 Flash, has raised concerns among copyright holders and users alike after it was discovered that the model can remove watermarks from images without permission. The tool, available in Google’s developer-facing tools like AI Studio, allows users to generate and edit image content with ease.
The Gemini 2.0 Flash model is particularly skilled at removing watermarks and filling in gaps created by their deletion. However, it lacks clear guidelines on its use and has been deemed “not for production use.” Despite this, the model’s performance has sparked debate among AI experts, who argue that some watermark removal capabilities are inherently unethical.
According to U.S. copyright law, removing a watermark without the original owner’s consent is considered illegal. In contrast, other models like Anthropic’s Claude 3.7 Sonnet and OpenAI’s GPT-4o explicitly refuse to remove watermarks due to concerns about potential illegality.
Google has yet to respond to requests for comment on this matter.
Source: https://techcrunch.com/2025/03/16/people-are-using-googles-new-ai-model-to-remove-watermarks-from-images