Google lets users turn off visible AI watermarks

Google said Friday it will let users remove the visible watermark from its AI-generated images, videos, and songs, though the change won’t affect the invisible SynthID watermark or the C2PA metadata standard embedded in the same files. Josh Woodward, Google’s vice president for Gemini, said in a post on X that the toggle will apply to the Nano Banana, Omni, and Lyria models, with the option to turn off the visible watermark available in Gemini and Google’s video editor, Flow — support for Search is coming soon.

“We’re striking a balance here between creative control and safety: while the visible watermarks are now optional, invisible SynthID watermarks and C2PA metadata are still being used for transparency. So you can still use Gemini or Search to see if an image was AI-generated,” Woodward wrote. The feature is rolling out over the coming days; once live, users will be able to toggle it under Settings > Media Watermark.

The change reflects a broader shift in how companies label AI-generated media: visible watermarks can make content less useful for professional and creative work, even as the need to identify AI-generated content remains. Google is also open-sourcing a new library called Credentio to let developers embed a local validation mechanism directly in their own apps. The move follows Anthropic’s hotly debated decision to include a watermark in text and files generated by Claude, in order to comply with EU regulations.

Why it matters

The shift shows Google leaning on invisible provenance signals rather than visible marks to satisfy transparency rules — a model other AI companies, including Anthropic, are also testing as regulators like the EU push for clearer AI-content labeling.