Google’s popular Nano Banana image generator has won praise for its creative output, but one persistent gripe from users was the visible watermark stamped on every generated image. Unlike images from ChatGPT’s latest model, which come without a built-in mark, Nano Banana’s creations always featured a small logo in the bottom right corner. That’s now changing. Google has quietly introduced a new setting that lets you toggle off the watermark, giving you cleaner, more professional-looking results.
The update is rolling out across Gemini and Flow, with Google Search’s AI Mode expected to follow soon. But the feature isn’t limited to images. You’ll also be able to strip watermarks from videos created with Gemini Omni and songs produced by Lyria. This move reflects Google’s growing understanding that creators want flexibility and a polished final product, especially when they’re using AI for client work or personal projects.
Why the Watermark Was a Problem
Watermarks have long been a standard way for AI companies to signal that content was machine-generated. They help maintain transparency and prevent misuse. However, for many users, the watermark felt like an unnecessary blemish, especially when the AI output was meant for professional use. Designers, marketers, and social media managers often had to crop or edit the corner of images, which sometimes ruined the composition. With the new setting, that hassle disappears.
The timing is also significant. As AI image generation becomes more mainstream, users are comparing tools side by side. ChatGPT’s image generator, for example, doesn’t add a visible watermark, which gave it an edge in certain workflows. By offering an opt-out, Google levels the playing field and makes Gemini a more attractive option for those who need clean, watermark-free assets.
How to Remove the Watermark in Gemini
Enabling the new feature is straightforward. In Gemini, head to the settings menu, usually accessible via your profile icon or a gear icon in the top right corner. Look for a section labeled “Generated content” or “Watermark settings.” There, you’ll find a toggle for “Remove watermark from generated images.” Flip it on, and any new images you generate will no longer carry the visible mark.
It’s important to note that this setting applies to future generations only. Existing images with watermarks won’t be retroactively cleaned. So if you have a library of previously generated images, you’ll need to regenerate them if you want watermark-free versions.
What About Flow and AI Mode?
Flow, Google’s newer collaborative workspace tool, also includes the same toggle. If you’re using Flow to brainstorm or create visual content, you can enable the watermark removal from the same settings panel. The company says the feature will also arrive in Google Search’s AI Mode, which lets you generate images directly from search results. That rollout is expected in the coming weeks, so keep an eye on your search settings.
Beyond Images: Watermark Removal for Videos and Music
Google’s watermark removal isn’t just for static pictures. Gemini Omni, the company’s video generation model, also benefits from this update. Videos created with Omni often carried a small watermark in the corner, which could be distracting in presentations or social media clips. Now, you can generate clean videos without any visible branding.
Similarly, Lyria, Google’s AI music generation tool, will let you remove watermarks from songs. This is a big deal for musicians and content creators who use AI to draft tracks or background scores. A watermark on an audio file is less visually intrusive, but it can still affect the listening experience if it’s an audible tone or a repeated announcement. Google hasn’t specified exactly how the Lyria watermark manifests, but the option to remove it will be available in the same settings area.
What About Metadata and Invisible Watermarks?
While you can now remove the visible watermark, Google is likely to retain invisible metadata tags. These are embedded in the file’s metadata and can be used to verify the content’s origin. This is a common practice among AI companies. For example, OpenAI includes C2PA metadata in images generated by ChatGPT, even though there’s no visible watermark. Google is expected to follow a similar approach, ensuring that the content can still be traced back to AI generation if needed.
For most users, this is a win-win. You get clean visuals without any distracting marks, while Google maintains a level of transparency behind the scenes. If you’re concerned about privacy or ownership, it’s worth checking Google’s official documentation on how metadata is handled.
How This Compares to Other AI Generators
ChatGPT’s image generator has never had a visible watermark, which made it a favorite for quick, clean images. However, it does embed invisible metadata. Midjourney, another popular tool, used to add a subtle grid pattern to images, but that was removed for paid users. DALL-E, also from OpenAI, includes a colored block watermark that can be removed with a paid subscription.
Google’s new setting puts it on par with these competitors, if not ahead. By giving users the choice, Google acknowledges that different users have different needs. A hobbyist might not mind a watermark, but a professional designer certainly does. This flexibility is a smart move in a crowded market.
Potential Downsides to Consider
Removing watermarks could raise concerns about misinformation. If anyone can generate realistic images without a visible label, it might be harder to spot AI-generated content online. However, invisible metadata and robust detection tools can help mitigate this risk. Google has also been developing SynthID, a tool that embeds a hidden watermark in AI-generated content. Even if you remove the visible mark, SynthID remains intact, allowing Google to identify the content as AI-generated.
So, while the visible watermark is gone, the underlying safety measures are still in place. This should give users peace of mind and help maintain trust in AI-generated media.
How to Get the Most Out of the New Setting
If you’re a frequent user of Gemini, here are a few tips to make the most of this update:
- Update your app: Make sure you’re running the latest version of Gemini or Flow. The setting may not appear on older versions.
- Check your default settings: If you generate a lot of images, consider setting the watermark removal as your default. This saves you from toggling it every time.
- Regenerate old images: If you have important images with watermarks, regenerate them with the new setting enabled.
- Test with videos: Try generating a short video with Gemini Omni to see how the watermark removal works in practice.
- Monitor AI Mode: Keep an eye on Google Search’s AI Mode. Once the feature rolls out, you’ll be able to generate watermark-free images directly from search.
What This Means for Creators and Businesses
For creators, the removal of watermarks is a game-changer. It means you can use AI-generated images in client presentations, marketing materials, and even print without worrying about unsightly logos. It also saves time, as you no longer need to manually edit out the watermark, which often involved cropping or cloning.
Businesses that rely on rapid content creation will also benefit. Social media managers can generate on-brand visuals without the need for post-processing. E-commerce sites can use AI images for product mockups, and bloggers can illustrate their posts with clean, professional images. The flexibility to choose whether to include a watermark is a thoughtful touch that puts the power back in the user’s hands.
The Future of AI Watermarking
This update raises an interesting question: Will other AI companies follow suit? As AI-generated content becomes more ubiquitous, the debate over watermarking is likely to intensify. Some argue that visible watermarks are essential for transparency, while others see them as a hindrance to creative work. Google’s decision to offer an opt-out suggests a middle ground—one that balances user convenience with responsible AI use.
In the coming months, we might see more tools adopt similar settings. For now, Google is leading the way by giving users control. Whether you’re a casual user or a professional creator, the ability to remove watermarks from Gemini-generated content is a welcome addition.
So, if you’ve been holding back on using Gemini because of the watermark, now is the time to give it another try. Head to your settings, flip the toggle, and enjoy clean, watermark-free AI creations. With this update, Google is making it easier than ever to integrate AI into your workflow without compromising on quality.
