이미지: TechCrunch AI
Summary
- Google is adding a setting in Gemini and Flow that lets users turn off the visible watermarks stamped on images, videos, and music
- Even with the setting off, the invisible SynthID watermark and C2PA metadata remain in place
- Google has also open-sourced a local verification library for developers called Credentio
- 발표일
- 2026년 8월 14일(금)
- 적용 모델
- Nano Banana, Omni, Lyria
- 적용 위치
- Gemini, Flow(구글 영상 편집 도구), Search는 추후 지원
- 설정 경로
- Settings > Media Watermark
- 미적용 대상
- 보이지 않는 SynthID 워터마크, C2PA 메타데이터
Only the visible mark gets switched off
Google is letting users turn off the visible watermark that has automatically been stamped on images, videos, and music it generates. Josh Woodward, VP of Gemini, said in a post on X that the toggle applies to the Nano Banana (image), Omni, and Lyria (music) models. The setting lives in the Gemini app and Google's video editing tool Flow, with support for Search coming later. The feature is rolling out over the coming days, and can be switched on or off via Settings > Media Watermark.
However, the invisible SynthID watermark and C2PA standard metadata remain in place regardless of this setting. Woodward offered a brief explanation, saying the change "balances creative freedom with safety." The same day, Google also released an open-source library called Credentio, which lets developers build their own verification logic into their apps.
Why there are two layers of watermarking
SynthID is a watermarking technology built by Google DeepMind that embeds patterns imperceptible to the human eye or ear directly into pixels or audio waveforms. This mark largely survives cropping or compression, so feeding a file into Gemini or Search can reveal whether it was AI-generated. The visible, logo-style watermark that has appeared in the corner of the screen until now serves a different purpose. It functions more as branding and an immediate warning label than as an identification tool, and as such has been an obstacle for users who want to put photos or videos directly to practical use.
Labeling AI-generated content has become a broader industry issue recently. Anthropic began adding watermarks to text and files created by Claude in order to comply with EU regulations, a move that drew pushback over concerns it could expose AI use in workplaces or schools. Google's latest move looks like the opposite approach, but is similar in substance. It leaves the visible marking to user discretion while keeping the identification functions that society needs — such as combating deepfakes or tracing content provenance — running at an invisible layer. C2PA is an international standard that cryptographically attaches a record of capture and editing history to photo and video files, and has been adopted jointly by companies including Adobe and Microsoft.
So what actually changes
Users who want to use images made with Nano Banana in ad creative or portfolios no longer need a separate editing step to remove the logo in the corner of the screen. But the fact that the file was AI-generated doesn't disappear. Running a SynthID check or looking up the C2PA metadata will still reveal the generation history, meaning that while the surface appearance becomes more natural, traceability remains intact. The open-sourcing of Credentio also opens a path for other developers to build similar verification logic directly into their own apps.
