Google eases visible watermarking for AI-generated content

Josh Woodward, vice president for Gemini, confirmed the update covers the Nano Banana, Omni, and Lyria models. The change addresses persistent feedback from professional creators who found visible markers intrusive for high-quality creative work. Users will find the new control under Settings > Media Watermark as the feature rolls out over the coming days, with integration for Google Search expected shortly.

To support broader industry standards, Google is also open-sourcing Credentio. This new library provides developers with the tools to build local validation mechanisms directly into their own applications. This shift follows competitive moves in the sector, such as Anthropic’s recent decision to mandate watermarking for Claude-generated outputs to align with evolving European Union regulatory requirements.