What happened?
Google has introduced the option to disable the visible watermark on images, videos, and music created with Gemini and its AI video generation tool, Flow. Users can turn off the new "Media watermark" setting in Gemini's settings to remove the "sparkle" mark that appears in the bottom-right corner of content generated with the Nano Banana and Omni models.
According to Josh Woodward, Vice President in charge of Gemini and AI Studio at Google Labs, while the visible watermark is now optional, the invisible SynthID watermark and C2PA metadata will continue to be embedded in the content behind the scenes. This allows users to check whether content was AI-generated through Gemini or Google Search.
Why does it matter?
Making visible watermarks removable could make it harder to quickly identify AI-generated content. However, many users were already removing these watermarks through editing anyway; this change essentially formalizes that practice.
Other major AI companies like OpenAI and Meta also don't add visible watermarks to their images. OpenAI relies on SynthID and C2PA standards, while Meta uses its own Content Seal standard. Anthropic, meanwhile, announced it has started applying invisible watermarks to both text and image content.
Companies' watermarking approaches
- Google: Visible watermark is optional, invisible SynthID and C2PA are always active.
- OpenAI: No visible watermark, verification provided via SynthID and C2PA.
- Meta: No visible watermark, uses its own Content Seal standard.
- Anthropic: Has started applying invisible watermarks to text and images.
What's next?
Google plans to bring the new watermark setting to Search as well. However, this disabling option won't be offered in countries that mandate the use of visible watermarks.