Google is rolling out a new option over the next few days that lets users hide visible watermarks on media created with its Gemini chatbot, in the standalone Gemini and Google Flow apps. The change applies to all images, videos, and songs generated with Gemini, except in countries where keeping watermarks is required by law, according to Josh Woodward, the vice president of Google Labs, Gemini, and AI Studio.
In the Gemini app, users toggle off the new "Media watermark" setting under Settings, a change that also applies to Google's AI video generator, Flow. When disabled, Google removes the "sparkle" watermark that appears in the bottom-right corner of content generated with its Nano Banana and Omni models.
Though visible watermarks are now optional, AI-generated content will still carry invisible SynthID watermarks and C2PA metadata embedded in the background, per Woodward. That means users can verify whether an image carries one of these markers by asking Gemini or Search if it's AI-generated.
SynthID embeds imperceptible signals that survive screenshots, crops, and format conversions, and tools like the SynthID Detector can confirm whether media came from an AI system.
Google also plans to roll out the new watermark setting to Search, but the toggle will not launch in countries that require visible watermarks. The change comes as other AI image generators, including OpenAI's and Meta's, already skip visible watermarks, with OpenAI relying on SynthID and C2PA and Meta using its own Content Seal standard.
