Google is making the visible Gemini watermark optional. The company said on Friday that users will be able to remove the visible mark stamped onto its AI generations — images, video and songs alike — through a new setting, with the change rolling out in the coming days.
Where the Gemini watermark setting lives
Once the rollout reaches your account, the control sits at Settings > Media Watermark, where you can turn the visible mark on or off. Josh Woodward, Google’s VP for Gemini, said in a post on X that the toggle covers the Nano Banana, Omni and Lyria models, and that it will work in Gemini and in Google’s video editor, Flow, with support in Search coming later.
Google is also open sourcing a library called Credentio, aimed at letting developers build a local validation mechanism into their own apps — a way to check a file’s provenance without routing it through Google.
The label you can see is the one going away
Google’s argument is that nothing about detection actually changes. The invisible SynthID watermark and the C2PA metadata standard both stay attached to whatever Gemini produces, regardless of the toggle.
“We’re striking a balance here between creative control and safety: while the visible watermarks are now optional, invisible SynthID watermarks and C2PA metadata are still being used for transparency. So you can still use Gemini or Search to see if an image was AI-generated.”
— Josh Woodward, Google VP for Gemini, in a post on X, as reported by TechCrunch
That is Woodward’s framing, and on a technical level it holds up. In practice it is worth being blunt about what the trade actually is. The corner watermark was the only AI label an ordinary person could see — on a screenshot, in a group chat, in a reposted video with the metadata stripped out by whatever app it passed through. SynthID and C2PA are real, but reading either one requires deliberately running the file through a checking tool. Almost nobody does that before believing a photo.
So the honest summary is that AI-generated media gets harder to spot at a glance and stays exactly as detectable for anyone who goes looking with the right software. Creators who found the mark intrusive in professional work get a genuine win. Everyone scrolling a feed loses the one cue that required no effort at all.
Phandroid reports that content will continue to carry the visible Gemini watermark in regions where the law requires AI labelling, so the toggle’s reach will depend on where you are.
What to do about it
Nothing, if you would rather keep the mark — the setting exists in both directions, and leaving it on preserves the current behaviour. If you generate images or video in Gemini for work and the corner badge has been getting in the way, check Settings for the Media Watermark entry over the next few days; staged rollouts of this kind usually take a week or more to reach everyone. And if you are on the receiving end of AI media rather than making it, the practical change is that you should trust your eyes slightly less than you did last week.
For more on how Google has been expanding what Gemini touches, see our coverage of Gemini’s 13 new connected apps.
Sources: TechCrunch, Phandroid
