Google Introduces AI Watermark Toggle in Gemini Ecosystem

Google has introduced an AI watermark toggle feature in its Gemini ecosystem, allowing users to disable visible watermarks on AI-generated content. The move aims to balance creative control for end-users with digital provenance safety.

The toggle option will apply across outputs including images, videos, and music generated by Google's Nano Banana, Omni, and Lyria AI models. According to Josh Woodward, VP for Gemini, the feature is designed to strike a pragmatic balance between creative control and digital provenance safety.

While users can turn off the visible watermark via media settings in Gemini and Google's video editor, Flow, the underlying security frameworks remain fully active. The invisible SynthID watermarking and Coalition for Content Provenance and Authenticity (C2PA) open-standard metadata will continue to embed into all media generations.

Google is also open-sourcing a developer library named Credentio to support localised metadata validation in external applications. This follows a move by competitor Anthropic, which recently introduced mandatory invisible text watermarking across its Claude platform to align with regulatory requirements under Article 50 of the European Union AI Act.