OpenAI Developing Music AI, Partners With Juilliard

OpenAI is developing a generative music tool that can produce music from text and audio prompts and has enlisted Juilliard students to annotate scores as training data. The company envisions uses such as adding guitar accompaniment to vocal tracks, generating background scores for videos, and enabling creators, advertisers and educators to produce music more easily. It is unclear when the tool will launch or whether it will be a standalone product or integrated into ChatGPT or OpenAI’s video app Sora, and OpenAI has not publicly detailed its progress. The effort builds on OpenAI’s earlier music research (MuseNet, Jukebox). Other firms including Google (Lyria), Suno and ElevenLabs are also advancing generative-music models while the rise of these tools has prompted scrutiny over copyright and training-data practices.