Google allows Gemini users to toggle off visible AI watermarks

Google Gemini artificial intelligence mobile application interface displayed on screen featuring vibrant colorful logo for productivity and assistant tasks. — Image credit: Search Engine Journal

Google is rolling out a new feature allowing users of its Gemini and Flow platforms to disable visible watermarks on AI-generated images, videos, and music, while still embedding invisible digital identifiers.

The move provides creators with more control over the aesthetic presentation of their AI-generated content, according to the company, while maintaining transparency through embedded metadata and imperceptible watermarks.

The new ‘Media Watermark‘ setting will apply to content generated by three specific models: Nano Banana for images, Omni for video, and Lyria for music. The feature is expected to become available to users globally over the next few days.

Despite the option to remove visible indicators, all AI-generated content will continue to carry invisible SynthID watermarks and C2PA metadata. These embedded identifiers are designed to signal that the content was created by artificial intelligence.

Josh Woodward, vice president of Google Labs, indicated that Google Search is slated to support these transparency measures in the future. He did not provide a specific timeline for this integration.

To further aid transparency and verification, Google has open-sourced Credentio, a C++ library. This library enables developers to directly verify C2PA credentials within their own applications, facilitating broader adoption of content authentication.

The option to turn off visible watermarks will not be available in certain countries where visible watermarks are legally mandated. These regions include India, South Korea, and Vietnam, specifically for video content.

Source: Search Engine Journal