📝 TL;DR 📝 Google rolled out a new Media Watermark setting on August 14 that lets Gemini and Flow users turn off the visible sparkle icon stamped on AI-generated images, videos, and music. It's a real, welcome quality-of-life fix for anyone doing professional or client-facing creative work. But it's worth being clear about what actually changed: the invisible SynthID signal embedded in the pixels or audio, and the C2PA provenance metadata baked into the file itself, stay in place regardless of what you choose. This removes visible branding, not the ability to verify a file's AI origin. 🧠 Overview 🧠 This is a genuinely useful update, and it's also a good, concrete example of a distinction worth understanding clearly: visible watermarks and invisible provenance signals solve two completely different problems, and removing one doesn't touch the other. Google's own VP for Gemini, Josh Woodward, confirmed the change directly on X, and notably framed it as responding to actual user demand, removing the visible watermark was one of the top ten most requested changes when Google asked users about their biggest Gemini complaints. The practical shift here is real: content made with Nano Banana (Google's image model), Omni (video), and Lyria (music) can now render without the visible sparkle badge, in the Gemini app and Google's Flow video editor today, with support coming to Search soon. 📜 The Announcement 📜 The toggle lives under Settings, Media Watermark, in both the Gemini app and gemini.google.com. Switching "Show watermark" off applies going forward to all new images, videos, and music you generate, it's not retroactive, and doesn't strip watermarks from anything you've already created. The rollout is gradual, reaching accounts over the next several days, so it may not appear in your settings immediately even after this announcement.