- Published on
Google Gemini to add option to disable visible watermarks on AI-generated images, videos, and music
- Authors
- Name
- DeepHorus
Google announced on August 14 via a post on X that it will add a "Media Watermark" setting for its AI assistant Gemini in the coming days, allowing users to turn off visible watermarks on AI-generated images, videos, and music. This feature can be found in the "Settings" menu in Gemini's left sidebar, positioned as the third item from the bottom, offering both "On" and "Off" options.
Previously, all content generated by Gemini included visible watermarks to indicate its AI-generated nature. As AI-generated content becomes more widespread, striking a balance between protecting creators' rights and enhancing user experience has become a focus of industry attention. Google's adjustment is a response to this challenge.
When "Media Watermark" is enabled, visible watermarks will remain on Gemini-generated content; when disabled, these visible watermarks are removed. However, Google emphasizes that even with the visible watermark turned off, AI-generated works will still retain SynthID invisible watermarks and C2PA metadata. SynthID is an AI content recognition technology developed by Google that can mark AI-generated content unobtrusively; C2PA metadata records the content's origin and editing information, providing a basis for content traceability.
This change means that users can still trace the AI-generated traces of content through technical means even after disabling the visible watermark. Google's move aims to balance transparency and visual appeal, satisfying users' aesthetic needs while not abandoning accountability for AI content. For content creators and platforms, the presence of invisible watermarks and metadata helps prevent misuse and maintain copyright order.
Industry analysts believe that this option from Google may be part of a broader trend in AI content management. With the increasing prevalence of AI generation tools, how to avoid misleading users while reducing watermark interference with the visual experience has become a common challenge for tech companies. Google's SynthID technology has previously been applied to multiple products, and its extension to Gemini-generated content now demonstrates its continued investment in AI content authentication systems.
Currently, this feature has not been officially rolled out; Google only stated that it will be launched in the "coming days." Users will need to wait for a Gemini update to use this setting. Whether other AI service providers will adopt similar practices remains to be seen. However, it is foreseeable that watermark strategies for AI content will continue to evolve between transparency and user experience.
