Google Allows Removal of Visible AI-Generated Watermarks
Photo: TechCrunch
Quick answer
Google now permits users to remove visible watermarks from AI-generated content (images, videos, audio) while retaining hidden SynthID tags and C2PA metadata for identification.
Google has announced a new feature allowing users to remove visible watermarks from AI-generated content, including images, videos, and audio. This applies to content created by models such as Nano Banana, Omni, and Lyria. However, hidden SynthID tags and C2PA metadata remain intact, ensuring content origin transparency.
According to Josh Woodward, Google's Vice President of Gemini, the new setting will be available in Gemini services and the Flow video editor, with plans to integrate it into search soon. Users can enable or disable visible watermarks in the media settings section. Woodward emphasized that Google aims to balance creative freedom with security while maintaining transparency through hidden technologies.
Google also unveiled the open-source Credentio library, enabling developers to integrate local authenticity verification mechanisms into their applications. This follows Anthropic's move to add watermarks to text and file generations in Claude to comply with EU regulations.
Common questions
- Why did Google allow the removal of visible AI watermarks?
- Google seeks to balance user creative freedom with the need for AI content identification. Visible watermarks can hinder professional use, but hidden tags preserve transparency.
- Which technologies does Google use to identify AI content?
- Google employs hidden SynthID watermarks and C2PA metadata standards, enabling AI content detection even after visible watermarks are removed.
- Where can users disable visible watermarks?
- The setting is available in Gemini services and the Flow editor, with future expansion to Google Search. The option is located in media settings.
Dzen feed: /feed/dzen.xml · RSS: /feed.xml