Google Made Gemini's Visible AI Watermark Optional — Here's Why That Matters

News

If you've gotten into the habit of looking for the small sparkle icon in the corner of an image to tell whether it came from Google's AI, that habit just got less reliable. Google has added a setting that lets people turn that visible mark off.

What changed. Gemini (and Google's video/music tool, Flow) now has a "Media watermark" switch in settings — On or Off. With it off, images made with Gemini's image tool, videos made with its video tool, and music made with its music tool no longer carry the small visible AI mark in the corner. You'll find it in the web version of Gemini by clicking the gear icon in the bottom left, then "Media watermark." Google has said it plans to bring the same option to Search.

What did NOT change. Turning the switch off only removes the mark you can see. Every file still carries an invisible digital watermark (called SynthID) and metadata (called C2PA) embedded in the file itself, regardless of the setting. Tools built to check for that hidden marker — like Google Search's "About this image" feature — can still identify the content as AI-generated even with the visible mark switched off.

One exception. In places where the law requires a visible AI label — China is the main example — Gemini will keep the visible watermark on no matter what the user's setting says.

Why this matters for you. Until now, "no sparkle in the corner" was a reasonable, if imperfect, sign that an image probably wasn't from Gemini. That's no longer true — anyone can now choose to hide it. If you're trying to judge whether a photo, video, or song you're looking at is real, a missing visible watermark tells you nothing on its own. Our guide on how to spot AI-generated photos covers what still works: a reverse image search and checking for that invisible Content Credentials (C2PA) label at a site like verify.contentauthenticity.org, both of which look past whatever visible mark a creator chose to show or hide.

Sources

Published August 17, 2026← All news
Radim S.
Founder & editor

Radim is a software developer who spends his days building with AI and his evenings explaining it to family members who don’t care how it works — only what it can do for them. The safety guides are checked claim by claim against primary sources before they go out.