Google DeepMind has developed SynthID, a tool to watermark AI-generated images such as deepfakes. Now in the beta phase, the technology adds a digital watermark inside image pixels, visible for ID purposes but imperceptible to the human eye. - After a White House meeting in July, Google and six other AI companies committed to watermarking audio and visual content to indicate that it is AI-generated. - Google is the first to offer SynthID, which provides three confidence levels for #Artificial Intelligence image detection, "Digital watermark detected," "Digital watermark not detected," and "Digital watermark possibly detected." - According to Google, SynthID can't be removed by simple editing or compression and should not compromise image quality. - SynthID will initially launch to a select group of business customers of Vertex AI, Google's cloud-based machine learning platform. - It's only available through Imagen, Google Cloud's own text-to-image model. https://www.technologyrev..