Gemini lets you turn off visible AI watermarks

AI watermarks are necessary for safety, but they have always been an annoyance for anyone actually trying to use the output in real projects.

Google is now rolling out a new toggle in the Gemini app that lets users turn off visible watermarks on AI-generated images, videos, and music. Users can simply flip the "Show watermark" switch in settings to apply their preference to all future creations.

Invisible watermarks stay put

Turning off the visible badge doesn't mean the creations are untraceable. Google's invisible SynthID watermarks and C2PA metadata will always remain embedded in the background regardless of your choice.

According to Google, this applies across images made with Nano Banana, videos from Omni, and tracks from Lyria. The feature is available in Gemini and Flow, with Search coming next, though visible watermarks will remain mandatory in regions where local laws require them. Users can also ask Gemini directly whether a piece of media was AI-generated.

Why it matters

For creators and marketers, this removes an everyday annoyance. Instead of cropping edges or using cleanup tools before publishing, you get clean, ready-to-use assets straight from the prompt.

We think Google found a smart balance here. Forcing visible stamps usually just drives creators toward third-party watermark removers. By letting the surface stay clean while embedding tamper-resistant SynthID and C2PA markers underneath, Google gives users the polished aesthetic they want without ditching traceability.