‘Inaudible’ watermark could identify AI-generated voices
The growing ease with which anyone can create convincing audio in someone else’s voice has a lot of people on edge, and rightly so. Resemble AI’s proposal for watermarking generated speech may not fix it in one, but
