Nvidia developed Fugatto, an AI model creating and altering music and audio. It generates sounds from text and modifies existing audio. The model can transform piano music into singing. It can also change accents and moods in speech. Nvidia is concerned about potential misuse. The company has not set a public release date.