Google has launched SynthID Detector, a web portal to help identify AI-generated content created with Google’s AI tools
SynthID Detector can detect invisible SynthID watermarks embedded in images, audio, video, and soon text produced by Google AI models
The portal provides a simple workflow: users upload content, the system checks for watermarks, and results highlight likely watermarked regions
The technology initially focused on images but now covers text, audio, and video, including outputs from Veo, Lyria, Imagen, and Gemini
The portal is currently available to early testers, with plans for broader access; researchers and journalists can join a waitlist
Google is building an ecosystem around SynthID, partnering with companies like NVIDIA and GetReal Security for broader watermark detection and verification
SynthID Detector aims to provide transparency and combat misinformation by making it easier to verify the authenticity of AI-generated media