< Back to all clusters
[TECHNOLOGY] · 65 sources

started · updated

Google releases SynthID Detector for public AI content verification

Google has released its SynthID Detector tool to the general public, making it freely available worldwide to identify AI-generated or edited images, videos, and audio. Previously, the tool was restricted to media professionals and researchers, but it is now accessible via synthid.com. Users must sign in using a Google, OpenAI, or Apple account to use the service.

The detector works by identifying invisible digital watermarks embedded in media. While it is designed to be imperceptible to humans, it can remain detectable even after common edits like cropping, compression, or filtering. Currently, the tool supports content from Google, OpenAI, NVIDIA, and Kakao, with support for Apple Intelligence expected to be added soon.

Google notes that the tool has limitations: it only detects SynthID-compatible watermarks and cannot identify all AI-generated content. Furthermore, the absence of a watermark does not prove a file is authentic. Since its inception in 2023, Google has applied SynthID watermarks to over 180 billion images and videos and approximately 240,000 years of audio content.

Entities

Apple · Google · Google DeepMind · Kakao · Nvidia · OpenAI · SynthID

Claims

What the coverage asserts, and how many sources carry each claim.

Sources