Google makes visible AI watermarks in generated media optional

Google is introducing a setting that lets users remove visible watermarks from AI generated images, videos and music created in Gemini and Flow. Emma Roth reports for The Verge that the option controls the small sparkle symbol shown in the lower right corner of media produced with Google’s Nano Banana and Omni models.

Users can switch off the new “Media watermark” setting in Gemini’s settings menu. Google plans to bring the control to Search as well. It will not be available in countries where regulations require visible watermarks on AI generated material.

Removing the on screen marker does not remove all provenance information. Josh Woodward, vice president of Google Labs, Gemini and AI Studio, says the company will continue to embed invisible SynthID watermarks and C2PA metadata in generated files.

SynthID is Google’s system for marking AI generated content in a way that is not normally visible to viewers. C2PA metadata provides information about a file’s origin and history. According to Woodward, users can check for these signals through Gemini or Google Search to determine whether content was made with AI.

Visible labels become less common

The change makes it harder to identify AI generated media at a glance. Yet visible watermarks have limited value when users can crop or edit them out. Google’s approach now more closely resembles that of other major AI providers. OpenAI generally uses SynthID and C2PA rather than visible watermarks, while Meta has introduced its Content Seal standard. Anthropic has also said it is adding invisible watermarks to AI generated text.