Google warned users not to misuse its AI tools in order to infringe copyright, reported TechCrunch [File]
| Photo Credit: REUTERS

Google’s Gemini 2.0 Flash model is being used to remove watermarks from copyrighted content, according to users on social media who tried out the feature, reported TechCrunch.

Users apparently discovered the trick when experimenting with Gemini 2.0 Flash’s image generation abilities. Some reported success with removing watermarks that are meant to stop the unauthorised re-use of protected images, while others said that some watermarks were only partially obscured, per the outlet.

However, a Gemini watermark was later applied to the image after it was morphed.

Watermarks are often used as paywalls on images, to make sure that they are not stolen or reposted without credit. Google’s Gemini Flash 2.0 model was able to remove watermarks in some cases and fill in their background to make these spaces match the rest of the image.

Image generation models are meant to be calibrated in order to prevent users from being able to execute such requests. However, content filters can sometimes fail or not work as expected due to the experimental nature of the technology.

Google warned users not to misuse its AI tools in order to infringe copyright, reported TechCrunch.