Google lets Gemini users turn off visible AI watermarks

In short: Google added a setting that lets people remove the visible watermark from AI-made images, videos, and music in Gemini and Flow.

What happened

Google says you can now toggle off a new setting called “Media watermark” in Gemini and in Google’s AI video tool, Flow. When the setting is off, Gemini will no longer add the small “sparkle” mark that shows up in the bottom-right corner of some AI-made media.

This applies to content created with Google’s Nano Banana and Omni AI models. A model is the system that generates the image, video, or music, like a recipe that tells the tool how to make the final result.

Even when the visible mark is removed, Google says it will still add hidden markers in the background. These include SynthID (an invisible stamp) and C2PA metadata (a set of built-in “content credits,” like a label attached inside the file rather than printed on top of it). According to Google executive Josh Woodward, people can check for these markers by asking Gemini or using Google Search.

Google also plans to bring this watermark toggle to Search. The company says it will not launch the toggle in countries that require visible watermarks.

Why it matters

Visible watermarks make it easier to spot AI-made media at a glance. If more people turn them off, it may become harder for everyday viewers to quickly tell what was generated by AI, especially on social media, where context is often missing.

At the same time, Google points out that many users already edit out visible watermarks, and several other AI tools do not use visible watermarks either. This update shows a broader shift toward hidden labeling that you have to check for, rather than an on-screen mark you can see immediately.

Source: The Verge AI