Google Now Lets Users Remove the Visible Watermark From AI-Generated Media

Google is making a notable change to how it labels AI-generated content. The company now lets users turn off the visible watermark that appears on AI-created images, videos, and songs produced by its Gemini models.

What's Changing

The new toggle applies to Google's Nano Banana, Omni, and Lyria models. Users will be able to switch the visible mark off through Settings, under the Media Watermark option. The control is rolling out first in the Gemini app and Google's video editor Flow, with Search support coming later.

Why It Matters

Visible watermarks have been a point of friction for creators. While they signal that content is AI-generated, they can also get in the way of professional and creative work. Google's Josh Woodward, VP for Gemini, framed it as a balance between giving users creative control and keeping transparency in place.

Invisible Watermarks Remain

Importantly, turning off the visible mark doesn't remove the invisible safeguards. Google's SynthID watermark and C2PA metadata still get embedded, so tools like Gemini and Search can still detect whether a piece of media was AI-generated.

Google is also open-sourcing a new library called Credentio, aimed at helping developers embed a local validation mechanism in their own apps.

The Bigger Picture

This move lands at an interesting time. Anthropic recently made headlines by adding watermarks to text and files generated by Claude, partly to comply with EU regulations. Google's approach — making the visible mark optional while keeping the invisible one — shows a different philosophy on how to label AI content without hampering everyday use.