Google Makes Visible AI Watermarks Optional in Gemini and Flow

Google is introducing an option that lets users remove visible watermarks from AI-generated images, videos, and songs. The change applies to content created with the company’s AI models, while invisible SynthID watermarks and C2PA-related metadata will remain in place.

The update gives users control over whether a visible mark appears on their generated media. It does not remove the underlying systems Google uses to help identify AI-generated content.

Visible Watermark Controls for Google AI Content

The optional visible watermark setting will be available for content generated through Google’s Nano Banana, Omni, and Lyria models.

Users will be able to turn the visible watermark on or off through the following setting:

• Settings > Media Watermark

Google said the feature will roll out over the coming days. Once it is available, the setting will appear in Gemini and Flow, Google’s video editor. Support for Search is also expected soon.

SynthID and C2PA Metadata Will Remain

Removing a visible watermark will not affect Google’s invisible SynthID watermark or metadata related to the C2PA standard.

That distinction matters. The visible label can be disabled, but Google will continue using the invisible watermark and C2PA metadata for transparency around AI-generated media. Google said Gemini or Search can still be used to check whether an image was created with AI.

The company described the update as a balance between creative control and safety. Visible watermarks can make generated content less useful for professional and creative work, while identifying AI-generated material remains part of the company’s approach.

Where the Google Watermark Toggle Will Be Available

Google’s visible watermark control is being introduced across selected AI models and products.

The setting applies to AI generations that include images, videos, and songs. Google has not said that the invisible SynthID watermark or C2PA-related metadata can be switched off.

Google Open Sources the Credentio Library

Alongside the watermark change, Google is open sourcing a new library called Credentio.

Credentio is intended to let developers add a local validation mechanism to their apps. The library relates to C2PA content credentials, which are part of the metadata Google says will continue to accompany AI-generated media even when a visible watermark is disabled.

AI Watermarking Remains an Active Industry Issue

Google’s announcement follows Anthropic’s move to include watermarks in text and files generated by Claude to comply with EU regulations. That move prompted debate among some Claude users.

Google’s update takes a different approach for its AI-generated media: the visible watermark is optional, while invisible SynthID watermarks and C2PA metadata remain in use.