Google's Gemini app is finally responding to one of the most persistent complaints from its users: the automatic visible watermark stamped on every piece of AI-generated content. A new setting called Media Watermarks will allow users to turn visible watermarks on or off across all Gemini creation tools, including images made with Nano Banana, videos generated with Omni, and music produced with Lyria. The update is rolling out over the next few days, and it marks a significant shift in how the company handles transparency in AI-generated media.
Until now, the visible watermark was applied automatically with no user control. Many creators and casual users found it intrusive, especially after spending time perfecting a generated image or composing a music track. Google has heard those complaints and is taking a more flexible approach, giving users the ability to decide whether their creations carry a visible marker. However, this does not mean that AI-generated content will become untraceable. Invisible SynthID watermarks and C2PA metadata will continue to be embedded in the background of every file, ensuring that the origin of AI-generated content remains identifiable through technical means even when no visible label appears.
How to turn off AI watermarks for Nano Banana, Omni, and Lyria in Gemini
The process for disabling visible watermarks is designed to be straightforward, in line with Google's usual consumer-friendly settings structure. Users need to open Gemini's settings menu, locate the option labeled Media Watermark, and toggle it off. Once the toggle is turned off, it applies to all future image, video, and music creations generated through the app. There is no way to retroactively remove watermarks from content that was created before the setting was changed, so users who want watermark-free outputs must remember to flip the switch before generating something new.
Google's own social media announcement described the rollout as happening over the coming days, meaning not every user will see the setting immediately. The company said that users can decide if their image, video, and music creations made with Gemini will have visible watermarks, while noting that invisible SynthID watermarks and C2PA metadata will always stay embedded in the background. That dual-layer approach is central to Google's strategy for maintaining responsible AI practices while also improving user experience.
What remains in the background: SynthID and C2PA metadata
Even with the visible watermark removed, Google ensures that provenance information is still attached to each file. SynthID is Google DeepMind's watermarking tool, designed to embed a hidden digital signature directly into the pixels or audio frequencies of generated media. This invisible marker is resilient to common transformations like cropping, resizing, compression, and even some forms of editing. It is designed to allow Google to identify whether content was produced by its AI systems, even if the visible label is absent.
In addition to SynthID, Gemini also attaches C2PA metadata to generated content. The Coalition for Content Provenance and Authenticity (C2PA) is an open standard supported by major industry players, including Adobe, Microsoft, Intel, and Google. C2PA metadata records information about the origin and editing history of a digital file, creating a tamper-evident chain of custody. This metadata can include the tool used to generate the content, the time of creation, and details about any subsequent edits. When users disable visible watermarks, they are only removing the human-readable label. The underlying C2PA record remains both readable by software and useful for verification platforms.
This layered approach reflects a broader industry trend. As generative AI tools become more powerful and widely available, governments, media organizations, and tech companies have been searching for ways to distinguish authentic content from AI-generated material. Visible watermarks are easy for ordinary users to spot, but they can be cropped out or covered up. Invisible watermarks and metadata offer a more robust method of tracking provenance, although they are not foolproof either. By combining both visible and invisible mechanisms, Google is attempting to balance user control with accountability.
Regional and account restrictions
Not all users will be able to take advantage of the new watermark toggle. Google has placed geographic restrictions on the rollout, and the reasons for those restrictions appear to be tied to local regulatory requirements and policy considerations. In India, South Korea, and Vietnam, the option to disable visible watermarks is only available to subscribers of the paid AI Ultra plan. These countries have distinct legal and regulatory environments around AI-generated content, misinformation, and platform accountability. Google may be taking a more cautious approach in those markets, requiring a premium subscription before allowing users to remove the visible marker. For free users in these countries, visible watermarks remain mandatory on all AI-generated images, videos, and music.
Additionally, work and school accounts do not receive the watermark toggle at all. For users who access Gemini through a Google Workspace for Education or a business enterprise account, visible watermarks will continue to appear on every generated asset. This is likely a deliberate decision by Google to maintain stricter oversight in institutional contexts. Organizations often use AI tools for internal projects, marketing materials, and educational content, and the presence of visible watermarks may help prevent unauthorized use or accidental distribution of AI-generated media. It also gives system administrators fewer variables to manage when ensuring compliance with internal policies or external regulations.
The distinction between consumer and institutional accounts is notable. Google has increasingly tailored feature availability based on account type. Consumer accounts often receive experimental or flexible options first, while enterprise and educational accounts are subjected to stricter controls. This approach allows Google to offer the latest conveniences to individual users while maintaining governance frameworks for organizations that may have legal obligations or brand safety concerns.
Why did Google add the AI watermark toggle to Gemini?
The decision to introduce this toggle did not happen in isolation. According to reports, Google VP Josh Woodward directly asked users about their biggest frustrations with Gemini. The response was overwhelming, with thousands of users submitting feedback. The request to remove Nano Banana's visible watermark quickly rose into the top 10 most commonly mentioned issues. Woodward acknowledged the feedback and said the team would consider making changes. The arrival of the Media Watermarks setting is the direct result of that community engagement process.
This development also comes amid growing community efforts to strip AI watermarks from generated content altogether. Some users have developed third-party tools and manual techniques to remove visible watermarks from images and videos created by various AI platforms. Others have shared workflows that involve cropping, inpainting, or using image editing software to erase labels. Google's responsive approach may be seen as an attempt to bring watermark control back into the official product rather than forcing users to resort to unofficial methods that could violate terms of service or compromise file integrity.
However, Google's decision is not purely a concession to user demand. By allowing users to turn off visible watermarks while keeping invisible ones, Google is modeling a more nuanced form of transparency. A visible watermark can degrade the aesthetic quality of an image or video, especially for professional creatives who may need clean visuals for client work. But the same users do not necessarily want to distribute AI-generated content without any trace of its origin. Invisible watermarks provide a way to preserve authenticity without interfering with the visual experience. For Google, this is both a user experience improvement and a strategic move that keeps the company aligned with emerging standards for responsible AI deployment.
The timing of the rollout is also relevant. Governments around the world have been considering legislation related to AI content labeling, deepfakes, and digital provenance. The European Union's AI Act, for example, includes transparency requirements for AI-generated content. While Google's watermarking policies predate the final approvals of many regulations, the company is actively preparing for a future where provenance is expected but not necessarily visible. Having built-in mechanisms like SynthID and C2PA gives Google a flexible toolkit that can adapt to different regulatory environments without requiring major changes to the user interface.
For creators, the new toggle is likely to be welcomed. Photographers, graphic designers, video producers, and musicians have increasingly been experimenting with generative AI tools. A visible watermark can interfere with a professional portfolio, a client presentation, or a social media post. Being able to turn off that watermark means creators can present their work in a cleaner format while still complying with ethical obligations through invisible metadata. Educational institutions and news organizations, on the other hand, may choose to keep visible watermarks on as a default to make it easier for their audiences to identify AI-assisted content.
There is also a philosophical question embedded in this update. If AI-generated content no longer carries a visible watermark, will users become less distrustful of what they see online? Some advocates for mandatory visible labeling argue that hidden metadata is not enough to protect consumers, especially those who do not know how to verify provenance information. Google's approach suggests that the company believes visible watermarks are not always the best solution. Instead, a combination of user-controlled visible labels and robust invisible markers may offer a more sustainable path forward. This is particularly important in a world where AI-generated content is becoming increasingly realistic and difficult to distinguish from human-created media.
As the rollout expands over the coming days, Gemini users will need to check their app settings to see if the Media Watermark option is available in their region and on their account type. For those who have access, the change offers a simple way to take control of how AI-generated content is presented. For those who do not, the policy remains unchanged: visible watermarks stay, at least for now. The update reflects a broader tension between transparency and usability, and Google's answer to that tension is a nuanced one that allows users to choose while ensuring that technical provenance remains intact. The company will likely continue to refine this feature based on feedback, just as it has done with the Gemini app overall, and it is easy to imagine further adjustments to regional availability, account policies, or watermarking defaults in the future.
Source:Digital Trends News