Social media users have discovered a controversial application of Google's new AI model, Gemini 2.0 Flash. Apparently, it is capable of removing watermarks from images, including those from Getty Images, Shutterstock, and other popular photo stocks. This was reported by TechCrunch.
Gemini 2.0 Flash causes copyright concerns
Last week, Google expanded access to the image creation feature in the Gemini 2.0 Flash model, which allows not only creating, but also editing graphic content. However, this feature seems to have weak protection against misuse.
Gemini 2.0 Flash does not hesitate to create images of celebrities and copyrighted characters and, as already mentioned, to remove watermarks.
Image by X user George Arrowsmith
As noted by X and Reddit users, the model not only removes watermarks, but also tries to fill the gaps left by the removal.
How people are doing it?
The process causes so much concern due to how easy it is to remove watermarks. All the user has to do is select Gemini 2.0 Flash (Image Generation) Experimental option in the model select, upload a picture to the chatbot and enter a prompt similar to "remove watermark from image." The AI responds to the request almost instantly.
However, it does not always do this correctly: it has difficulty with translucent marks and watermarks that cover a large part of the image. Below is an example where the word 'Getty' was replaced with 'Geck' during processing.
Image by X user Malhar Ujawane
Other AI-based tools have similar capabilities, but Gemini 2.0 Flash is particularly adept at this and is also available for free.
Nevertheless, AI models by Google's competitors, such as Anthropic's Claude 3.7 Sonnet and OpenAI's GPT-4o, refuse to remove watermarks, claiming it is a copyright violation to do so without the owner's consent.
As a reminder, Google has expanded access to Deep Research in Gemini for all users and introduced a new experimental feature with personalization.
