started · updated
Google launches SynthID Detector to identify AI-generated content
Google has released the SynthID Detector, an online tool designed to help users identify whether images, videos, or audio files have been created or manipulated using artificial intelligence. The tool works by scanning uploaded content for specific, imperceptible digital watermarks embedded by Google and certain partners.
Users can access the service via SynthID.com by uploading a file and receiving a simple yes or no response regarding the presence of the watermark. The detector is compatible with content generated by Google’s Gemini models, as well as models from partners such as OpenAI, NVIDIA, and Kakao. However, support for Apple-generated files is still under development.
The tool has notable limitations. It can only detect content that utilizes SynthID technology; therefore, it cannot identify AI-generated media from models that do not use this specific watermarking method, such as those from Meta. Additionally, the detector indicates whether a watermark is present but does not specify which exact parts of a file were AI-generated. Usage is subject to a limit of approximately 10 verifications per user per day, and a login via a Google, OpenAI, or Apple account is required.
Entities
Gemini · Google · Nvidia · OpenAI · SynthID Detector