OpenAI Reportedly Developing AI Music Generator From Prompts
Juilliard students annotate scores as trials test video scoring and vocal backing; ChatGPT could host the tool.
Summary
- OpenAI is reportedly developing an advanced AI music-generation tool that creates full tracks from text and audio prompts
- The company is collaborating with Juilliard students to annotate musical scores to improve the AI’s training data
- The tool’s emergence in the tense AI music market raises key questions about licensing, copyright, and artist compensation
OpenAI is reportedly developing an AI music-generation tool that turns text and audio prompts into full tracks. Early use cases include scoring videos and adding accompaniment to vocals. Multiple outlets also cite trial work with Juilliard students annotating scores to improve training data.
There’s no official confirmation or launch date. The tool could live inside ChatGPT or tie into OpenAI’s fast-growing video stack, where sound is already a focus. With Sora 2 headlining OpenAI’s push into video-plus-audio creation, a music generator is a logical next play but still leaves big questions around licensing, consent and revenue sharing.
The industry backdrop is tense. AI music leaders face lawsuits over training data. If OpenAI enters, expect pressure to ink label deals and enforce guardrails to avoid style cloning. For creators, a polished tool could mean instant, bespoke cues, ad jingles and demo beds without expensive libraries or studio time. For artists, it intensifies debates about attribution, compensation and the limits of mimicry.
If realized, the move would pull music into OpenAI’s end-to-end creative pipeline across text, images, video and sound, raising the stakes for the whole generative arts ecosystem.
OpenAI has a track record in music AI. The company previously shipped MuseNet (2019) and Jukebox (2020) — “a neural net that generates music, including rudimentary singing”.
