Truth that Matters. Stories that Impact

Truth that Matters. Stories that Impact

Technology

Google to Allow Users to Turn Off Visible Watermarks on AI-Generated Media

Google has announced an update allowing users to remove visible watermarks from content generated by its artificial intelligence tools, covering images, audio tracks, and video clips. While the visible tags can be switched off, Google confirmed that its invisible identification systems will remain embedded in the files to preserve transparency.

What Happened

Josh Woodward, Vice President for Gemini at Google, shared details of the policy shift on social platform X. According to Woodward, users will soon be able to disable visible watermarks on content produced by Google’s Nano Banana, Omni, and Lyria AI models. The option will be introduced across Gemini and the video editing tool Flow, with integration for Google Search planned to arrive shortly afterward.

Despite the visual change, Google confirmed that underlying provenance measures will not be altered. Output files will continue to carry invisible SynthID watermarks along with Coalition for Content Provenance and Authenticity (C2PA) standard metadata. Users will still be able to use Gemini or Search to detect whether an image was created using artificial intelligence.

Key Highlights

  • Optional Visible Tags: Users can turn off visible marks on AI-created images, videos, and songs.
  • Supported Models: The feature applies to generations from the Nano Banana, Omni, and Lyria models.
  • Unchanged Invisible Tracking: Invisible SynthID watermarking and C2PA standard metadata remain fully active.
  • Platform Availability: The setting is launching in Gemini and the Flow video editor, with Search support scheduled to follow.
  • Simple Toggle: Once rolled out, users can navigate to Settings > Media Watermark to adjust the toggle.
  • Open-Source Tooling: Google is releasing an open-source library named Credentio, enabling third-party developers to integrate local watermark validation into their own applications.

Why This Matters

Visible watermarks on AI-generated media often limit their practical use in creative and professional workflows. By making visible badges optional while preserving invisible digital signatures, Google aims to provide flexibility for creators without sacrificing the ability to identify synthetic media. This development follows recent industry debates regarding content labeling, including Anthropic’s implementation of watermarks in Claude-generated files to meet European Union regulatory requirements.

What to Watch Next

The feature will roll out to users over the coming days. Following its release on Gemini and Flow, functionality will expand to Google Search. Developers can also look out for the open-source Credentio library to implement local verification mechanisms in their respective software.

Frequently Asked Questions

Which AI models support the removal of visible watermarks?

The toggle will be available for media created using Google’s Nano Banana, Omni, and Lyria models.

How can users turn off visible watermarks?

Once the update reaches their account, users can access the control by going to Settings > Media Watermark within supported applications.

Does turning off the visible watermark remove all AI identification?

No. Google confirmed that invisible SynthID watermarks and standard C2PA metadata will continue to be embedded, allowing tools like Search and Gemini to identify the content as AI-generated.

Source: TechCrunch, reporting on a statement by Josh Woodward on X.