Google lets users remove AI visible watermarks
Google announced users can remove visible watermarks from AI images, videos, and songs, while keeping invisible SynthID intact.
Stock photo for illustration only, not from the actual event
- Google allows users to remove visible watermarks from AI-generated images, videos, and songs.
- Invisible SynthID watermarks and C2PA metadata remain active to ensure content transparency.
- The toggle feature applies to Nano Banana, Omni, and Lyria models via settings.
- The company is open-sourcing a new Credentio library for local developer validation.
Google announced on Friday that it will now grant users the capability to strip away visible watermarks from media generated by its artificial intelligence systems, encompassing images, videos, and songs. The tech corporation specified that this update will leave the invisible SynthID watermarks and metadata tied to the C2PA standard completely unaffected.
Josh Woodward, the company's vice president for Gemini, revealed through a post on the platform X that this toggle switch will be rolled out for the Nano Banana, Omni, and Lyria models. He detailed that the setting to disable the visible watermark will be integrated directly into Gemini and Google's video editing software, Flow, with integration for Search scheduled to follow shortly.
Stock photo for illustration only, not from the actual event
This adjustment mirrors an evolving strategy regarding how AI-generated media is labeled, noting that visible watermarks frequently render digital content less practical for professional assignments and creative projects, even though identifying artificial creations remains a necessity.
"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"
Writing about the initiative, Woodward explained that they are maintaining equilibrium between creative freedom and safety by making visible markers optional while utilizing invisible SynthID tags and C2PA metadata for continuous oversight, allowing verification via Gemini or Search.
Google's shift highlights the ongoing friction between safety labeling and professional utility, as prominent visual stamps often hinder commercial adoption of generative media. Retaining backend verification via SynthID offers a balanced approach that preserves aesthetic freedom while satisfying auditability demands.
The feature is currently rolling out over the next few days, allowing users to navigate to Settings followed by Media Watermark to toggle the mark on or off. Additionally, Google is moving to open-source a library named Credentio, designed to enable software developers to embed a local validation framework directly into their own applications.
This rollout follows closely behind Anthropic's much-debated decision to implement watermarks inside text and files produced by Claude in order to adhere strictly to European Union regulations.
Source: TechCrunch
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment