The Information reports that OpenAI is working on a new tool to generate music based on text and voice prompts.
Sources say such tools could be used to add music to existing videos or add guitar accompaniment to existing vocal tracks. It's unclear when OpenAI plans to release such a tool or whether it will be available as a standalone product (rather than an integration with OpenAI's ChatGPT or video app Sora).
A source told The Information that OpenAI is working with some students at the Juilliard School to annotate scores as a way to provide training data.
OpenAI has announced generative music models in the past, but those predate ChatGPT's announcement. More recently, the company has developed audio models focused on text-to-speech and speech-to-text conversion. Other companies with generative music models include Google and Suno.
TechCrunch has reached out to OpenAI for comment.