Firstly, watermarking isn't just a Google thing
When you generate an image using Google's Nano Banana model in Gemini, you'll have noticed the small sparkle watermark that gets added to the corner. Up until recently, there was no way to turn it off. That's changed.
Watermarking on AI generated content isn't something exclusive to Google and Gemini. It's standard practice across AI image tools generally, not just Nano Banana. Google's version just happens to be the one in the news right now, because you can finally switch the visible part off.
Google confirmed the update in mid-August 2026. There's a new setting called Media Watermark that lets you turn the visible sparkle icon on or off for anything you create in Gemini, including images from Nano Banana, video from Omni, and music from Lyria.
Google's VP for Gemini, Josh Woodward, said it's about balancing creative control with safety. The visible mark is optional now, but the invisible layers stay in place no matter what you choose.
3. Find Media Watermark - Toggle it OFF
You can turn it off in your account settings, only if this has been enabled for your region. Google announced the update in mid-August 2026, and the rollout is staged, so it may not have reached everyone yet and isn't showing up on mobile apps immediately either. India, South Korea, and Vietnam only get the option with a paid AI Ultra subscription, and work or school accounts don't get it at all, so the watermark stays mandatory there. Once you do turn it off, it only applies going forward. Anything you already generated keeps its visible watermark.
Switching off the visible watermark doesn't make an image untraceable. Google still embeds an invisible SynthID watermark and C2PA metadata into every file, whether the visible mark is showing or not. Google's own tools, including "About this image" and Lens, can still confirm an image came from Gemini regardless of your watermark setting.
If the toggle isn't available to you yet, or you're working with an image that already has a watermark on it:
Turning off the watermark changes how the image looks, not what it is. Google's SynthID and C2PA metadata are still there in the background, and Google's own detection tools can still tell it's AI generated either way.
We've written about this in more detail in our post on whether AI images and stock photos hurt your SEO. It's worth crediting that post here too, since a lot of what applies to AI images generally applies to Nano Banana images specifically. Rankings come down to how relevant, well-optimised, and useful an image is on the page, not whether it was AI generated.
AI tools are becoming part of everyday business quickly. You might already be using ChatGPT to help draft an email, come up with content ideas or answer a question but if you use Google, Gmail, Docs or Drive every day, there’s another AI tool you’re probably going to start seeing a lot more. Google Gemini. […]
Firstly, watermarking isn't just a Google thing When you generate an image using Google's Nano Banana model in Gemini, you'll have noticed the small sparkle watermark that gets added to the corner. Up until recently, there was no way to turn it off. That's changed. Watermarking on AI generated content isn't something exclusive to Google […]
Creating original images for every page of your website takes time, effort and money. That is why so many businesses use stock photography and AI-generated images. They’re fast, affordable and make getting a website or blog post live easier. Do stock photos hurt SEO? And do AI images impact SEO? Stock photos and AI images […]
