Google Will Let Users Remove Visible Watermarks From AI-Generated Images, Videos, and Songs
Google will soon let users switch off the visible watermark on its AI-generated images, videos, and songs, with a toggle rolling out across Gemini and the Flow video editor in the coming days and Search support following soon after. TechCrunch
Josh Woodward, Google's VP for Gemini, said the toggle will cover the Nano Banana, Omni, and Lyria models. Users can turn the visible watermark on or off via Settings > Media Watermark. The setting will be available in Gemini and Google's video editor Flow, with Search support coming soon.
The key detail: removing the visible watermark does not affect the invisible SynthID watermark or the C2PA metadata that Google embeds in its AI-generated outputs. SynthID is Google's system for adding an invisible digital watermark to AI-generated images, videos, and audio without reducing quality. A tool called SynthID Detector can check whether an image was created or edited using Google AI, and the Gemini app also uses SynthID to verify an image's origin.
Google is also open-sourcing a library called Credentio, which lets developers build a local C2PA content-credentials validation mechanism into their own applications. C2PA is an industry standard for attaching provenance information — essentially a verifiable chain of custody — to digital media. By open-sourcing Credentio, Google is making that verification accessible outside its own products.
The broader context here involves several threads coming together. At Google I/O 2026, Google announced an expansion of SynthID, and OpenAI disclosed that it is adopting the watermarking standard. The original commitment traces back to July 2023, when Google, OpenAI, and others pledged to watermark AI content for safety in a White House announcement. Google is also keeping the visible watermark — the Gemini sparkle — on images generated by free-tier and Google AI Pro-tier users of Nano Banana Pro, meaning the new toggle does not apply uniformly across all product tiers.
More recently, Anthropic announced it would include a watermark in text and files generated by Claude to comply with EU regulations. Google's decision to make its visible watermark optional moves in the opposite direction from Anthropic's compliance-driven approach, while leaning harder on the invisible SynthID layer as the primary provenance mechanism.
The tension is between user control and content authentication. Visible watermarks serve an immediate signaling function: someone scrolling through a feed can tell at a glance that an image was AI-generated. SynthID, by contrast, requires specialized tooling to detect. Ars Technica reported in late July that simply cropping SynthID-watermarked images can often eliminate the watermark, raising questions about how robust the invisible layer is in practice. If users remove the visible mark and the invisible mark is vulnerable to routine manipulation like cropping, the combined system has a meaningful gap.
The architectural bet Google is making is worth flagging. By open-sourcing Credentio and expanding SynthID's adoption — including OpenAI's integration — Google is positioning invisible, metadata-level provenance as the industry standard rather than visible labels. The C2PA standard provides a structured way to assert where content came from, and a local validation library lowers the barrier for third-party apps to participate. The question is whether invisible watermarks, however well-designed, can survive the real-world environment of image distribution, where crops, re-encodes, and platform-level stripping of metadata are routine.
Google's announcement also lands alongside a busy product cycle. The company announced Gemini 3.7 Flash, which it describes as its most intelligent workhorse model yet for coding and agents, and published its Made by Google 2026 collection of Pixel device announcements. The watermark toggle, Credentio, and model updates together suggest Google is consolidating its AI provenance and generation tooling into a more coherent platform.
The pragmatic read is that Google is acknowledging what has been true since visible watermarks were introduced: they are a friction point for creators who want to use AI-generated content in polished contexts, and they are trivially separable from the underlying asset in any case. Doubling down on SynthID and C2PA metadata while giving users control over the visible mark is a reasonable hedge, provided the invisible layer holds up under the kind of casual manipulation that Ars Technica's testing exposed. That provision is doing a lot of work, and the next round of independent SynthID robustness testing will tell us whether Google's two-layer approach is structurally sound or just more convenient.
