The world is now rife with fakes—and this is no longer just hyperbole from sappy 1980s song lyrics. Easy access to generative AI tools allows for the rapid, free creation of fake images, videos, and audio. This morning, Google opened access to SynthID Detector for users worldwide.
SynthID Detector identifies images, videos, and audio generated using AI. It provides a percentage-based estimate indicating the likelihood that the content is AI-generated. Google states that the tool supports AI content from several organizations, including OpenAI, NVIDIA, Kakao, and—in the future—Apple.
Google has long anticipated the issues that would arise from widespread public access to generative AI tools. Since 2023, they have implemented the SynthID system, which embeds invisible watermarks into AI-generated content. SynthID is now a standard feature in many generative AI tools on the market.
According to Google, SynthID has so far embedded watermarks into over 180 billion AI-generated images and videos, as well as 240,000 years' worth of AI-generated audio. SynthID Detector can be accessed via the link below.
