Some of the latest “experimental” features of Google’s Gemini 2.0 Flash model have raised serious concerns among many users. In particular, the model has demonstrated “very good” results in removing watermarks from images.
Gemini 2.0 Flash can not only generate images based on a text request but also edit them according to the user’s instructions — examples of this functionality have been shared on Reddit. Recently, it was discovered that the model can remove watermarks with high accuracy. This behavior is markedly different from OpenAI’s GPT-4o model, which actively rejects requests to remove watermarks.
Previously, there were tools such as Watermark Remover.io, which allow users to remove watermarks from stock photos. In addition, a Google research team created a similar algorithm back in 2017 to highlight the need for more reliable copyright protection for images. The difference now is that Gemini 2.0 Flash appears to outperform earlier tools, especially when dealing with complex watermarks.
Advanced Editing Capabilities in Gemini 2.0 Flash
Gemini 2.0 Flash seems particularly effective at removing intricate watermarks, such as the Getty Images stamps, and is capable of restoring the image with high quality. Naturally, after removing the watermark, the model adds a SynthID tag, effectively replacing the copyright label with “edited with AI.” However, AI tags like SynthID are relatively easy to remove using other AI tools, as Samsung’s object eraser tool has demonstrated in the past.
According to user reports, the lightweight Gemini 2.0 Flash model can also add recognizable images of real people to photos — a feature that the full Gemini model does not allow. This capability has further fueled concerns regarding possible misuse.
Currently, these advanced editing functions are accessible only to developers via AI Studio, notes NIX Solutions. The publicly available Gemini bot refuses to process requests involving the removal of security tags, as shown in the screenshots shared by users. Google has not responded to inquiries about whether the model includes safeguards against potential misuse once it becomes widely available.
We’ll keep you updated as more information about Gemini 2.0 Flash’s public release and possible protections becomes available.