Google lets Gemini users switch off the visible watermark on AI images
Archive · Issue of 21 August 2026

What happened
On Friday 14 August, Google added a Media Watermark setting to the Gemini app and its Flow video tool. It lets people turn off the visible sparkle mark on AI-made images (Nano Banana), video (Omni) and music (Lyria). The invisible SynthID watermark and C2PA provenance data stay in every file (Search Engine Journal, Google Gemini Help). Google’s Josh Woodward called it “striking a balance here between creative control and safety”.
Why it matters for your business
If you use Gemini for product shots, social posts or flyers, you can now publish without the corner logo. That looks cleaner. It also means nothing on the image tells a customer it was made by AI, unless you say so.
For most marketing images that is fine. It is not fine for an AI image of a real person, or a “photo” of a shop, product or event that never existed. That is where labelling rules bite, and where customers feel misled.
Anyone can still check a file. Gemini and Google’s “About this image” can read the hidden SynthID mark, so an unlabelled AI image can be identified later.
The Swiss angle
The setting sits under Settings, then Media Watermark in Gemini on the web. Google’s help page lists limits only for India, South Korea and Vietnam. Some press reports said the toggle would not appear in the EU; we could not confirm that from Google, so check your own account. Google does not say whether work (Workspace) accounts get the toggle.
The rules are clearer. Since 2 August 2026, the EU AI Act requires businesses that publish deepfake images, audio or video to disclose it (Jones Walker). That reaches Swiss firms whose content targets EU customers. At home, the Federal Data Protection Commissioner says manipulated faces, images or voices of identifiable people “must also always be clearly indicated” (FDPIC).
What to do
Make a one-line house rule: AI images of people, or of things that look like real photos of your business, get a visible “AI-generated” label, watermark or not. Everything else is your call.
Sources
- Search Engine Journal, 17 August 2026: Google Lets You Turn Off Visible Watermarks In Gemini
- Google Gemini Help: Manage watermark settings in Gemini Apps
- Dataconomy, 17 August 2026: Google Makes Visible AI Watermarks Optional In Gemini
- GIGAZINE, 17 August 2026: Google has added a toggle to remove visible watermarks
- Jones Walker: Yes, August 2 Still Matters
- FDPIC: AI and data protection
Legal note: This article is general information, not legal advice. Laws change and every business is different, so always get advice from a qualified lawyer or your fiduciary (Treuhänder) before you act on it.
Get the free newsletter
Every Friday: three AI stories that matter for Swiss businesses, and what to do about them. Free, in English, German, French or Italian.