What’s happened? As per an exclusive report by The Information, OpenAI has quietly begun work on an AI model that can generate music, taking its next big leap beyond text and video. The company is reportedly training its systems using annotated music data from Juilliard students, which is a strong hint that it’s chasing quality and creative precision rather than simple generative output. The move positions OpenAI to rival AI-music pioneers like Suno and Udio, even as those startups face major copyright lawsuits from the music industry.
- Internal discussions suggest the AI could create music using both text and audio prompts, similar to how its Sora tool creates videos from text.
- The project aims to support multiple use cases, from ad jingles and video background scores to full-length compositions, and will possibly be integrated into ChatGPT or Sora.
- This follows OpenAI’s previous, now-retired experiments in music AI: MuseNet (2019) and Jukebox (2020), which were precursors to today’s large multimodal models.
Why this is important: This isn’t just another feature, but the next battleground between creators and machines. By diving into music generation, OpenAI isn’t just aiming to stay ahead of rivals, but it’s also challenging who owns creativity and how we will consume it.
- Creators vs. AI: Start-ups like Suno are already facing lawsuits from record labels, claiming models were trained on copyrighted songs. OpenAI’s entry puts the big guns in the ring.
- Service-ecosystem play: OpenAI has over 800 million ChatGPT users; adding music means more reasons to stay in its ecosystem, and more hooks for monetisation.
- Legal and ethical pivot: With Sora’s deepfake controversy showing how far the guardrails still have to go, OpenAI’s music ambition will raise fresh questions about licensing, consent, and revenue-sharing.
Why should I care? AI-generated music might sound like a novelty today, but it’s quietly redefining how creative industries operate. From the way ads are made to how indie musicians record demos, tools like this could reshape workflows, royalties, and even what it means to “make” music. But as we’ve seen with Sora’s deepfake controversy, power without permission is risky, and the same playbook could repeat here if guardrails don’t keep up.
- If OpenAI’s music tool works as intended, anyone could compose soundtracks or jingles without needing expensive software or studio time.
- For artists and producers, it could become both a threat and a tool by offering faster creativity but also raising fears about AI “copying” signature styles.
- Consumers may get richer, AI-personalized audio in videos, games, and ads, but also face murky copyright lines and authenticity concerns.
Okay, so what’s next? One can expect OpenAI to push an announcement in late 2026 or 2027 outlining its music-generation tool, potentially integrated into ChatGPT or the Sora ecosystem. However, it won’t be smooth. In fact, one of the first things to watch is whether it secures licensing deals with major labels or whether it ends up in court alongside the likes of Suno. Simultaneously, the early backlash from Sora’s deepfakes illustrates that OpenAI’s guardrails are lagging behind its ambitions. As reported by Reuters, creator agencies like Creative Artists Agency are already warning the company about risk and rights. Safe to say, the beat won’t drop cleanly until the legal chaos and creator fury settle.
