Google Lets Some Gemini Users Turn Off Visible AI Watermarks
It would have been easy to crop this out, if you were okay with losing some of the lake.
Credit: Google
Google has started rolling out a setting that lets some Gemini users turn off visible watermarks on AI-generated images, videos, and music. The controversial change means these users can more easily pass off AI content as human-made.
The toggle appears under “Media Watermark” in the settings for Gemini, Nano Banana, Omni, and Lyria. When users turn it off, Gemini stops adding the visible mark to new creations; watermarks from past content are unaffected.
This setting only affects Google’s visible AI watermarks, which could otherwise be cropped out of visual content. Invisible SynthID watermarks will remain regardless of whether the toggle is on or off. C2PA metadata also stays, which can record information about a file’s origin and editing history.
The change could be Google’s way of making its AI content generators more competitive. Most people who generate images with AI want a clean image, not one with an obvious watermark, which could lead them to choose other free tools such as ChatGPT, Claude, or Copilot. Sure enough, when Gemini VP Josh Woodward asked X users last month to voice their complaints about Gemini, many indicated that they’d like the option to exclude visible watermarks from AI-generated images.
“We’re striking a balance here between creative control and safety,” Woodward said of the change on Friday.
Most Gemini users now have access to the toggle, with exceptions in India, South Korea, and Vietnam, where only AI Ultra subscribers have access. Work or school accounts that offer Gemini apps also exclude the toggle.
