Google Gemini gained a new “Media Watermark” control on Thursday, August 20, 2026, letting users turn off the visible watermark for AI-generated images, videos, and music—an upgrade creators have wanted since watermark overlays started getting more noticeable in decks and social posts.
According to Tom’s Guide, the walkthrough is designed so you can remove the visible layer without cropping the output or jumping into an external editor. The catch: it doesn’t erase every marker of AI origin, since Google says invisible transparency measures remain in place.
Google gemini: 2026-08-20: The Setting That Turns Off the Visible Watermark
The inciting moment happened on Thursday, August 20, 2026, when Tom’s Guide covered Google’s rollout of a new “Media Watermark” setting inside Gemini.
The setting is meant for creators who want their AI-generated assets to look less branded when they’re used in presentations, product mocks, or content calendars. Worth noting: the feature targets the visible watermark that Google normally adds—not the invisible tracking layer.
Here’s the thing: the change is positioned as an appearance-control toggle, not a “remove AI provenance” switch. Google also ties this control to its Flow filmmaking tool for video creation and says Google Search will get the same option later.
In other words, the UI work is happening across the Gemini ecosystem rather than being a one-off experiment.
Rolling Out Across Gemini and Flow (So Your Workflow Doesn’t Break)
That said, the rollout matters for how people actually use AI media tools day-to-day. If your workflow starts in Gemini, generates an image or video, and then drops it into PowerPoint-style slides, a visible watermark can become a distraction.
Google’s decision to make the option part of the creation flow means you get the “final-looking file” without extra editing steps.
The feature is also specific about coverage: it applies to images made with Gemini’s Nano Banana, videos created with Omni, and music generated with Lyria, per Tom’s Guide.
Those names are important because they map to different model pipelines, and the point is that the watermark control spans multiple media types. For developers and agencies, it also suggests Google is moving from “one watermark fits all” toward “user-controlled presentation layers with machine-readable transparency staying intact.”
Why It Still Doesn’t Remove Everything That Signals AI
In the early 2020s, watermarking became the battleground between two groups: creators who want clean aesthetics, and platforms that need accountability and traceability. Google is essentially choosing the middle path here by keeping the visible element optional while leaving its invisible mechanism untouched. For wider coverage, see The Verge.
Tom’s Guide notes that while the visible watermark can be disabled, Google will continue embedding its invisible SynthID watermark and C2PA content signals. This is where the debate gets real.
Critics may argue that “visible-only” control still leaves audiences susceptible to confusion in contexts like pitch decks and marketing thumbnails. Google’s rebuttal is that disabling the visible mark improves creator control, while the invisible layers still preserve transparency for systems and platforms that can read them.
So the practical takeaway is simple: if you’re trying to make a cleaner output for humans, you can. If you’re trying to eliminate all traces of AI generation from both people and systems, this isn’t that.
What’s Next: More Control, Then More Consistency Across Surfaces
Worth noting: Google says the setting is currently rolling out across Gemini and Flow, and Google Search will later receive the same option. That implies Google is trying to standardize how watermark controls behave across surfaces, so a “turn it off” moment doesn’t require re-learning different menus. For wider coverage, see TechCrunch.
It also sets up a future where AI media can look presentation-ready by default in certain contexts, but still carry behind-the-scenes provenance. For the India and global creator community, this could reduce last-minute formatting churn.
Agencies often resize, crop, and re-export assets specifically to hide watermark zones, and that’s time-consuming. If you’re building repeatable workflows around Gemini outputs, the “Media Watermark” toggle removes one recurring friction point—while leaving the platform’s transparency story intact.
That forward-looking move is why this matters beyond the UI: it’s a shift toward creator-grade finishing options inside the generation app itself, not after it.
Related Articles
- Made in India, Built to Beat the World: Murf AI's Falcon 2 Takes On OpenAI and ElevenLabs
FAQs
How do I remove watermarks in Google Gemini?
Open the new Media Watermark setting in Gemini and turn off the visible watermark option, as described by Tom’s Guide.
Does turning off the visible watermark remove all AI traces?
No. Google says the visible watermark can be disabled, but invisible transparency measures such as SynthID and C2PA remain embedded.
Is this feature only for images?
No. The setting applies to AI-generated images in Gemini, videos in Flow and related model pipelines, and music from Google’s model stack, with rollout across Gemini and Flow.
Will the option appear in Google Search too?
Google says the same watermark control will come to Google Search later, after the current Gemini and Flow rollout. Stay tuned for more on google gemini.
Was this article helpful?
Your feedback directly improves future articles on this site.
