Google Makes Visible AI Watermarks Optional in Gemini and Flow
Google will let Gemini and Flow users remove visible AI watermarks while retaining SynthID and C2PA data, with Search support and Credentio coming soon.
Summary
Google said on Friday, August 14, 2026, that users will be able to remove visible watermarks from AI generated images, videos, and songs. Gemini VP Josh Woodward said the toggle will cover the Nano Banana, Omni, and Lyria models in Gemini and Google’s Flow video editor, with Search support coming soon. The feature will roll out in the coming days through Settings, then Media Watermark.
Removing the visible mark will not affect Google’s invisible SynthID watermark or C2PA metadata. Woodward said users can still check through Gemini or Search whether an image was AI generated, preserving identification while giving professional and creative users cleaner outputs. Google is also open sourcing Credentio, a library that lets developers embed local validation in apps. The shift follows Anthropic’s hotly debated addition of watermarks to Claude generated text and files to comply with EU regulations.
Positives
- Nano Banana, Omni, and Lyria users will gain control over visible watermarks in Gemini and Flow.
- SynthID watermarks and C2PA metadata will remain attached, preserving methods for identifying Google’s AI generated media.
- Credentio will be open sourced so developers can add local validation mechanisms to their apps.
- Gemini and Search will continue helping users determine whether an image was AI generated.
- Optional visible marks can make Google generated images, videos, and songs more useful for professional and creative work.
Risks & concerns
- Optional visible watermarks reduce direct visual labeling of AI generated images, videos, and songs.
- Search support is only described as coming soon, with no specific release date provided.
- The rollout will take several days, so the Media Watermark setting will not reach every user immediately.
- Anthropic’s watermarks for Claude text and files, introduced for EU compliance, show that AI labeling remains hotly debated.