Google: Gemini's Visible Watermark Becomes Optional

Google activates a setting that removes the "sparkle" logo from content generated with Gemini and Flow. The invisible markings SynthID and C2PA metadata remain embedded, according to Josh Woodward. The button is also expected to arrive in Search, but will not be available in countries that require a visible watermark.

Google will not launch the button where a visible watermark is mandatory

Google plans to extend this setting to Search but will not enable the deactivation in countries where visible marking is required. In the ecosystem, other image generators like those from OpenAI and Meta already do not apply a visible watermark. OpenAI relies on SynthID and C2PA for marking, while Meta has launched its own standard, Content Seal. Anthropic, for its part, has announced that it applies invisible watermarks to texts and images generated by its systems.

In Gemini and Flow, a setting disables the "sparkle" watermark

A new "Media watermark" switch in Gemini and in Flow, Google's AI video tool, allows users to disable the display of the watermark. Once turned off, the "sparkle" logo in the bottom right disappears from content generated with the Nano Banana and Omni models. The option is accessible from the settings menu of Gemini. Images, videos, and music are affected by this change.

SynthID and C2PA remain in the background, according to Google

Even without visible marking, invisible SynthID watermarks and C2PA metadata continue to be embedded in generated content, according to Josh Woodward, vice president of Google Labs, Gemini, and AI Studio. It remains possible to verify if content is AI-generated by querying Gemini or Search. This development could complicate the quick identification of a visual or video, as many users were already editing content to remove visible logos.