Google lets users turn off Gemini AI visible watermarks
Users can now toggle off visible watermarks on images, videos, and music made with Gemini and Flow, while invisible SynthID and C2PA markers remain.
Stock photo for illustration only, not from the actual event
- Google now allows users to remove visible watermarks from AI-generated media.
- The feature is controlled via a Media watermark toggle in Gemini and Flow.
- The sparkle watermark in the bottom-right corner can now be switched off.
- Invisible SynthID and C2PA watermarks remain embedded for verification purposes.
Google has rolled out an update enabling users to remove visible watermarks from images, videos, and music generated through its artificial intelligence tools. With this new update, creators can easily toggle off a designated Media watermark setting within Gemini and Google's AI video generation platform, Flow.
When this specific setting is switched to off, Google will clear away the sparkle watermark that typically appears in the bottom-right section of content generated by the company's Nano Banana and Omni models, giving users clean outputs without the explicit AI branding overlay.
This shift by Google brings its platform policy closer to competitors like OpenAI and Meta, neither of which apply visible watermarks to their AI-generated images. Instead, the industry is increasingly leaning toward backend verification standards, balancing creative flexibility with provenance tracking.
Despite making visible watermarks completely optional, Josh Woodward, vice president of Google Labs, Gemini, and AI Studio, confirmed that AI-generated content will still feature invisible SynthID watermarks and C2PA metadata embedded directly into the background of the files.
This backend integration ensures that individuals can still verify whether a specific piece of media was created using artificial intelligence simply by asking Gemini or Google Search to check for these embedded markers, even if no visual indicator is present on the surface.
Looking at the broader industry landscape, other major AI developers have adopted similar strategies. OpenAI utilizes SynthID and C2PA to signal AI generation, Meta introduced its own Content Seal standard, and Anthropic announced this week that it is applying invisible watermarks to both AI-generated text and images.
Looking ahead, Google has outlined plans to expand this new watermark control setting to Search as well. However, the toggle will not be launched in countries and regions that maintain legal requirements compelling the permanent display of visible watermarks on AI-generated content.
Source: The Verge
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment