\n<\/p>\n
Audio created using Google DeepMind\u2019s AI Lyria model, such as tracks made with YouTube\u2019s new audio generation features<\/a>, will be watermarked with SynthID to let people identify their AI-generated origins after the fact. In a blog post<\/a>, DeepMind said the watermark shouldn\u2019t be detectable by the human ear and \u201cdoesn\u2019t compromise the listening experience,\u201d and added that it should still be detectable even if an audio track is compressed, sped up or down, or has extra noise added.<\/p>\n<\/div>\n According to DeepMind, SynthID\u2019s audio implementation works by \u201cconverting the audio wave into a two-dimensional visualization that shows how the spectrum of frequencies in a sound evolves over time.\u201d It claims the approach is \u201cunlike anything that exists today.\u201d<\/p>\n<\/div>\n The news that Google is embedding the watermarking feature into AI-generated audio comes just a few short months after the company released SynthID in beta<\/a> for images created by Imagen on Google Cloud\u2019s Vertex AI. The watermark is resistant to editing like cropping or resizing, although DeepMind cautioned that it\u2019s not foolproof against \u201cextreme image manipulations.<\/a>\u201d<\/p>\n<\/div>\n