Customize Gemini: Disable Visible Watermarks

Google is reaffirming its commitment to transparency in AI-generated media by continuing to embed invisible SynthID and C2PA watermarks within images, videos, and music created using its artificial intelligence tools. These unperceivable markers are designed to provide a persistent indicator of AI origination.

However, a significant new update will grant users the ability to remove visible watermarks from AI-generated content. This functionality is introduced through a new "Media watermark" setting that can be toggled off within Gemini and Google's AI video generator, Flow.

Disabling this setting will eliminate the distinctive "sparkle" watermark, typically located in the bottom-right corner of content produced by the company's Nano Banana and Omni models. Despite visible watermarks becoming optional, all AI-generated content will still carry invisible SynthID watermarks and C2PA metadata embedded in the background, as confirmed by Josh Woodward, Vice President of Google Labs, Gemini, and AI Studio. This ensures that the AI origin of an image can still be verified by querying Gemini or Search.

While this adjustment may slightly complicate the immediate identification of AI-generated images or videos, it acknowledges that many users have already been manually removing visible watermarks. Furthermore, Google's approach aligns with other major AI image generators, such as those from OpenAI and Meta, which do not employ visible watermarks. OpenAI, for instance, relies on SynthID and C2PA for AI content signaling, while Meta has developed its proprietary Content Seal standard. This week, Anthropic also announced its implementation of invisible watermarks for both AI-generated text and images.

Google intends to extend this new watermark setting to Search in the future. However, this toggle will not be available in regions where visible watermarks are mandated by local regulations.