Google announced on Friday that it will allow users to disable the visible watermark on AI-generated content produced by its systems, covering images, videos and songs. The company emphasized that this change does not remove the invisible SynthID watermark or the C2PA-standard metadata, which will remain in place to support provenance and transparency.
Josh Woodward, Vice President for Gemini, said in a post on X that the toggle to enable or disable the visible watermark will be available for the Nano Banana, Omni and Lyria models. He noted the setting will appear in Gemini and in Google’s video editor, Flow, with support for Search coming soon.
Google framed the update as an attempt to balance creative control and safety: visible watermarks can impede professional and creative use of generated assets, but identifying AI-generated content remains important. According to Woodward, while the visible watermark becomes optional, the invisible SynthID marks and C2PA metadata will continue to allow tools such as Gemini or Search to determine whether an image was produced by AI.
The feature is scheduled to roll out in the coming days. Once available, users will be able to toggle the visible mark on or off via Settings > Media Watermark.
In addition, Google open-sourced a new library called Credentio to let developers embed a local validation mechanism into their apps.
The announcement comes against a broader industry backdrop: Anthropic recently introduced a controversial watermark in text and files generated by Claude to address EU regulatory requirements. Google’s change can be seen as another approach to reconciling the needs of usability and provenance.
