Google Finally Lets You Remove Gemini's Visible AI Watermark

Google is making a small but very welcome change to Gemini. Users can now turn off the visible watermark that appears on AI-made images, videos and music. The option is rolling out now, and it means you can finally get a clean file straight from Gemini without having to crop or edit the watermark out later.

There is one important detail, though. Google is only removing the watermark you can see. Its hidden SynthID watermark and C2PA data will remain inside the file, so there will still be ways to check if the content was made with Google's AI tools.

The new option is called Media Watermark, and you can find it inside Gemini's Settings menu. It has a simple On or Off switch. Leave it on and Gemini will keep adding the usual visible mark. Turn it off and your new AI content will no longer have it.

And this isn't just for images. Google says the setting works with AI-made images, videos and music. It will be available for content made with Nano Banana, Gemini Omni and Lyria, while the option is also coming to Google's Flow video tool. Search support will follow later.

To be honest, I am surprised Google kept the visible watermark for this long. It was useful as a quick way to spot AI content, but it also made Gemini less useful for real creative work. If you're making an image for a presentation, website or social post, having a Gemini logo sitting in the corner isn't exactly ideal.

Here's the thing, though. Turning off Media Watermark does not make the file impossible to trace back to AI.

Google will still add its SynthID watermark in the background. Unlike the visible mark, SynthID is built into the content itself and is designed to survive common edits. Google also keeps C2PA metadata, which can store details about how a file was created or changed.

That feels like a better solution to me. The person making the content gets a clean image or video, while Google still has tools that can help check where it came from.

Google has been moving in this direction for some time. When Nano Banana Pro arrived last year, the company removed visible watermarks for AI Ultra users and developers working through AI Studio, while free and Pro users still got the Gemini mark.

There is a catch though. Google says the new switch will not be offered in countries where local rules require a visible watermark on AI content. For everyone else, the setting is rolling out over the next few days.

I think this is the right move. Anyone who really wanted to remove the old watermark could already crop it out or use an editing tool, so it wasn't doing much to stop AI content from being passed off as something else.

Keeping SynthID and C2PA in place makes far more sense. Gemini users get cleaner files, while the tools needed to check their origin remain in the background. That's a much better balance than putting a small AI logo on everything Gemini creates.