Google lets users Gemini remove visible watermark with new Media setting

Google has added a Media Watermark setting that lets users turn off Gemini’s visible watermark on generated content. The change applies in Gemini and Flow, and Google also plans to bring it to Search’s AI Mode. It gives creators a cleaner-looking output option while keeping Google’s hidden provenance systems in place.

Gemini watermark off option also applies to videos and songs

The new control is not limited to still images. Google says the Gemini remove visible watermark off option also works for videos made in Omni and music generated with Lyria. That means users can manage the visible stamp across more than one type of AI output without needing separate settings for each format.

For people working with AI-generated assets, that is the main change here. Instead of seeing a visible mark on every piece of content, they can choose a cleaner presentation when needed. Google is still keeping the feature tied to its own tools, so the watermark control sits inside Gemini and Flow rather than being a broad public editor.

SynthID watermark still remains with C2PA metadata in Gemini

Turning off the visible watermark does not mean Google is removing all identification signals. The hidden SynthID watermark still remains, and C2PA metadata in Gemini stays embedded in the content. In other words, the public-facing stamp can be reduced, but the underlying provenance layer is still there.

That distinction matters because it separates what people see from what Google can detect later. The visible watermark is the obvious label on the surface. SynthID and C2PA metadata work in the background, helping indicate whether the content was generated by AI. Google’s approach keeps those identity controls active even when the visible mark is switched off.

The company is also planning to extend the setting to Search’s AI Mode, which suggests the same watermark behavior may show up across more of its AI products. For now, the key point is simple: Gemini users can remove the visible watermark, but the content does not become anonymous. Google still keeps its detection and metadata systems attached.

That makes the new Media Watermark toggle a practical change rather than a full removal of safeguards. Users get more flexibility in how their AI-generated images, videos, and songs appear, while Google retains the tools it uses to track origin. For anyone comparing AI content workflows, this is a small but meaningful shift in how visible branding and hidden provenance are handled.