Google’s Gemini app is adding a Media Watermarks setting that lets users disable visible AI watermarks on generated media, according to Digital Trends. The report says the toggle covers images made with Nano Banana, videos made with Omni, and music made with Lyria. Until this change, Digital Trends reports, Gemini applied a visible watermark automatically with no user-facing option to turn it off. The new control sits in Gemini’s settings under Media Watermark, where users can switch visible watermarks on or off. The selection applies to future generated images, videos, and music tracks, rather than retroactively changing previously created files. The change is not a full removal of provenance signals. Digital Trends reports that Google still embeds an invisible SynthID watermark and Coalition for Content Provenance and Authenticity metadata into generated content regardless of the visible watermark setting. In practice, the update changes what appears on the surface of the media, not the underlying identification layer Google says remains attached behind the scenes. There are also access limits. Digital Trends says users in India, South Korea, and Vietnam only receive the option if they have a paid AI Ultra subscription; otherwise, their generated content keeps the visible watermark automatically. Work and school accounts do not get the setting, according to the report, so visible watermarks remain mandatory for those accounts. Digital Trends frames the rollout as a response to Gemini user feedback. The report says Google vice president Josh Woodward asked users to name their biggest Gemini frustrations, and that removing Nano Banana’s visible watermark landed among the top 10 requests. Woodward had said the team would consider it, and the option has now arrived, according to the outlet. For Google, the move separates two debates that often get blurred together: presentation and provenance. Users who dislike stamped visible labels get more control over the look of AI-generated media, while Google continues to preserve invisible attribution mechanisms. Based on the material in this cluster, the exact availability beyond the named restrictions and the timing of the full rollout remain single-source details. Who benefits: Gemini users who create images, videos, or music and want cleaner visible outputs benefit most, where the setting is available. Paid AI Ultra users in India, South Korea, and Vietnam are the only users in those countries reported to receive the option. Who's exposed: Users on work or school accounts are still exposed to mandatory visible watermarking, according to the report. Teams relying on Gemini output should not assume the visible watermark can be removed across all account types or countries.