Google has launched SynthID Detector, a new tool that helps identify AI-generated content across multiple media formats. This comes as advances in generative AI make it possible to create high-quality content, raising questions about authenticity and context.
The SynthID Detector portal provides quick detection capabilities for AI-generated content made with Google’s AI tools, including text, audio, images, and videos. It can also highlight areas of the content that may have been watermarked with SynthID.
To combat misinformation and misattribution, Google first introduced SynthID, a watermarking tool that preserves content quality while enabling detection even after sharing or transformations. The technology has already watermarked over 10 billion pieces of content.
The new detector works by scanning uploaded media for the SynthID watermark. For audio, it highlights specific segments where the watermark is detected, and for images, it indicates areas likely to be watermarked.
Google is starting to roll out SynthID Detector to early testers today, with a waitlist available for journalists, media professionals, and researchers. The company has also opened sourced its text watermarking technology and partnered with NVIDIA to expand the use of SynthID on their platform.
Additionally, Google is partnering with GetReal Security to detect watermarked content. This collaboration aims to provide more transparency tools for people engaging with AI-generated content, emphasizing the importance of continued collaboration between the AI community and industry partners.
Source: https://blog.google/technology/ai/google-synthid-ai-content-detector