Google has enabled users to disable the visible “sparkle” watermark on images, video, and music generated by its AI models. This change applies to content created via the Gemini app and the Flow video generator, removing the logo from the bottom-right corner of the media file. The update allows creators to toggle this visual marker off in the settings for both Nano Banana and Omni models. However, the removal of the visible mark does not strip the underlying digital identifiers. Invisible SynthID watermarks and C2PA metadata remain embedded within the background data of every file. Josh Woodward, vice president of Google Labs, confirmed that these hidden tags persist to maintain traceability even when the public-facing logo is absent. The move addresses user complaints about cluttered interfaces while attempting to balance visibility with practical usability. It represents a shift from mandatory branding to an optional overlay, yet the technical signature of the content remains intact for verification purposes.
This adjustment matters because it distinguishes between aesthetic branding and technical provenance. While the visible logo disappears, the file retains cryptographic proof of its origin. This ensures that platforms or individuals can still detect AI-generated material using specific detection tools. The company maintains control over the content’s identity without forcing a visual element on every output. Users retain full editorial control over the appearance of their work while accepting that the digital footprint remains.
* Visible “sparkle” logos are now toggleable off
* Invisible SynthID and C2PA metadata stay embedded
* Applies to Gemini, Flow, Nano Banana, and Omni models
