Google launches SynthID Detector to catch fraudsters

In response to the proliferation of counterfeits, misinformation, and fraud using artificial intelligence online and in educational institutions, Google DeepMind introduced SynthID Detector on Tuesday. This new tool scans images, audio, video, and text for invisible watermarks embedded in the growing set of Google artificial intelligence models.

SynthID Detector, designed to work with multiple formats in one place, aims to enhance transparency by identifying content created by Google's artificial intelligence, including AI audio NotebookLM, image generator Lyria, and Imagen, and highlighting fragments that are most likely to be watermarked.

"When working with text, SynthID determines which words will be generated next and adjusts the likelihood of selecting appropriate words that will not affect the overall quality and usefulness of the text," says Google’s demonstration video.