Google announced a new tool called SynthID Detector at Google I/O designed to identify AI-generated content created with Google’s AI technologies. According to Pushmeet Kohli from Google DeepMind, this verification portal can efficiently detect AI-generated images, text, audio, and videos by scanning for SynthID watermarks. These watermarks are embedded in content produced by Google’s models such as Gemini, Imagen, Lyria, and Veo. When users upload media, the detector highlights specific parts likely containing these watermarks. For audio, it pinpoints segments, and for images, it indicates watermarked regions. The tool is initially rolling out to early testers and will expand to those on a waitlist, allowing Google to enhance content transparency based on user feedback.
Back
