Google has announced that it will allow users to remove the visible watermark from content generated by its AI tools, including images, videos, and music. The company clarified that this change will not affect the invisible watermark known as SynthID, nor the metadata associated with the C2PA standard for authenticating content sources.

Google is enabling the removal of the visible watermark from AI-generated content.

The option includes images, videos, and songs generated via Google Forms.

The invisible SynthID watermark remains enabled to distinguish the content.

The feature is coming to the Gemini app and Flow editor, with search support coming soon.

The announcement came via a post by Josh Woodward, Vice President of Gemini at Google, on the X platform. Woodward explained that the move aims to balance giving users more creative freedom with maintaining security, as the visible watermark becomes optional while the invisible watermark and metadata remain for transparency purposes.

What changes to the watermark?

A watermark is a label added to content to indicate its source, and it comes in two types: visible, which the user can see on the image or video, and invisible, embedded within the file and difficult to detect with the naked eye. Google's change only applies to the visible type, leaving the invisible type active as is.

This move is part of an evolving approach to tagging AI-generated content, making the visible watermark optional after it was previously mandatory for all output. This need stems from the fact that the visible watermark sometimes reduces the suitability of content for professional and creative uses, while the need to distinguish generated content remains.

Woodward noted that the option to remove the watermark will be available for Google's generative models: Nano Banana for images, and Omni and Lyria for music generation. This gives users greater control over the final output across the various media types produced by these models.