Google Lets Users Turn Off AI Watermarks in Gemini and Flow

Google has quietly made a significant change to how AI-generated content is handled across its platforms. Users can now disable the visible "sparkle" watermark that previously appeared in the bottom-right corner of images, videos, and music created with Gemini and Flow, the company's AI video generator. The update, reported by Emma Roth at The Verge, introduces a new "Media watermark" toggle that gives creators direct control over whether that visual marker appears on their output.

It sounds like a minor UI tweak. For business content teams, it is anything but.

What Actually Changed

Previously, any content generated through Gemini or Google's Flow platform carried a visible watermark — a small but unmistakable signal that the material was AI-produced. The new toggle, accessible in settings for both Gemini and Flow, allows users to switch that marker off entirely.

Google has not removed its underlying metadata watermarking through SynthID, its invisible digital watermarking technology. That layer remains intact regardless of what you toggle in settings. What has changed is the visible, consumer-facing label that appeared on the finished asset.

This distinction matters. Google is giving users cleaner output for production use while retaining its own ability to verify AI provenance behind the scenes.

Why This Matters for Business Teams

For marketing departments, agencies, and small business owners who have been experimenting with AI-generated visuals and video, the visible watermark was a real friction point. Publishing a product image or a promotional video with a visible AI badge raises questions — about brand perception, about audience trust, and in some industries, about regulatory optics.

Removing that friction opens a practical door. Teams can now incorporate Gemini-generated visuals directly into client deliverables, social media campaigns, and internal presentations without the watermark creating an awkward conversation or requiring a post-production cleanup step.

There is also a competitive signal here. The ability to produce clean, professional-grade AI content without visible attribution is increasingly becoming table stakes. Adobe Firefly, Midjourney, and other platforms have already offered watermark-free output to paying subscribers. Google aligning with that standard — and doing so across both static and video formats — signals that it is positioning Gemini and Flow as serious tools for professional workflows, not just consumer experiments.

The Transparency Trade-Off

The move does introduce a genuine tension worth acknowledging. Visible watermarks, for all their awkwardness in finished content, served a transparency function. Audiences encountering AI-generated media in the wild had at least a fighting chance of identifying it. Without the visible marker, that signal disappears at the point of consumption.

Google's answer to this is SynthID. The invisible watermark persists in the file's metadata, theoretically allowing verification tools to confirm AI origin even when the visual badge is gone. In practice, however, very few consumers, journalists, or procurement teams have the tools or the habit of checking metadata before engaging with a piece of content.

For businesses, this creates a responsibility gap. Just because you can remove the watermark does not mean you always should. Regulated industries, editorial environments, and client relationships that depend on disclosure standards may still call for transparency — whether or not Google requires it. Having a clear internal content policy around disclosure is now more important than ever.

What SMBs Should Do Now

If your team uses Gemini or Flow for content production, this update deserves attention at a workflow level, not just a settings level. A few practical steps worth taking:

Review your content disclosure policy before toggling off watermarks across the board
Clarify with clients or stakeholders whether AI-generated content requires attribution in their context
Use the cleaner output to accelerate production, but document AI usage internally for compliance and audit purposes
Explore how AI tools for business are evolving around transparency standards, especially in verticals like finance, healthcare, and publishing

Teams building out structured AI workflows — where content is generated, reviewed, approved, and published at scale — should also think about how platforms like WRRK.ai can help centralize that process, keeping production moving without losing visibility into what was made, how, and when.

Original reporting by Emma Roth, published at The Verge on August 14, 2026.

Ready to build smarter AI workflows for your team? Explore WRRK.ai and get started today.

Frequently Asked Questions

Can Google still track AI-generated content after the watermark is turned off?

Yes. Google's SynthID technology embeds an invisible watermark in the metadata of AI-generated images, videos, and music. Turning off the visible "sparkle" watermark in Gemini or Flow settings does not remove this underlying layer. Google retains the ability to verify AI provenance even when the visual badge is not displayed.

Do businesses need to disclose that content was made with AI if there is no visible watermark?

This depends on your industry, jurisdiction, and client agreements. While Google no longer requires visible labeling, many regulatory frameworks and editorial standards still require disclosure of AI-generated content. Businesses in advertising, finance, healthcare, and publishing should review applicable guidelines and establish a clear internal policy regardless of what platform-level settings allow.

What is the difference between Gemini and Google Flow?

Google Gemini is Google's flagship AI platform, capable of generating text, images, and multimodal content. Google Flow is a dedicated AI video generation tool. Both platforms now support the new Media watermark toggle that allows users to disable the visible AI label on generated content.

Google Lets Users Turn Off AI Watermarks in Gemini and Flow — What It Means for Business Content Teams