There are many free image editing services on the web. Among the popular ones are to increase the size of the image or remove the watermark. Removing watermarks is often a topic of discussion because it seems like stealing the results of other people's images sold online. Most recently, Google did a brave job after OpenAI and Anthropic stated that they would not do it.
What is it? The Gemini 2.0 Flash (Image Generation) Experimental model is now available on the Google AI Studio development site. This model was developed specifically to help users edit images by simply uploading the image they want to edit, entering a short prompt and waiting about 10 seconds to 30 seconds for the final result to be generated. One of the best examples is to remove watermarks from images. Images with large, distracting watermarks or with clear text will be easy to remove and generate users with more or less the same image.
However, because this AI model is still in the development stage, it still does not have enough data to clearly understand the user's context. In addition, this model is also seen not to give users the same image because if the watermark is complex and disrupts part of it, this AI model will try to generate more or less similar content to fill the void left after the watermark is removed.
At the moment it is still uncertain when Google will include this model in the stable version of Gemini. It is also uncertain whether Google will limit the ability of Gemini 2.o Flash version to generate this image with a rule not to remove watermarks on the results of the real photographer's images. What do you think?