You may soon be able to generate music from your text prompts, audio prompts, or both with the new tool OpenAI is currently working on. For example, when the tool comes to live, you'll be able to “create upbeat jazz guitar for a cooking video” or “add string accompaniment to your vocal track”.
In addition to that, the tool may allow users to input text descriptions of mood, genre, instrumentation, input audio prompts—e.g., an existing vocal track, add music to existing videos, or melody—and then get a full musical output—such as instrumentation and accompaniment—accordingly.
This update comes few days after the AI company launched ChatGPT Atlas—an AI browser that features a ChatGPT sidebar within the browser that allows users to ask questions about the current webpage, summarize information, compare products, and analyze data from any site.
However, OpenAI is collaborating with students at the Juilliard School to annotate musical scores—supplying high-quality training data to build the model’s understanding of structure, harmony, and instrumentation.
The Juilliard School is a renowned performing arts conservatory in New York City, located at Lincoln Center. It was founded in 1905 and offers undergraduate and graduate degrees in dance, drama, and music.
The school also has a Preparatory Division for younger students, and the Tianjin Juilliard School in China offers graduate studies and other programs, as well as a global K-12 curriculum.
But it's important to note that the actual date of launching is unknown.
It’s also not yet known whether OpenAI Music will be a stand-alone product or app, a new feature inside ChatGPT, or integrated with OpenAI’s video tool, Sora — as opposed to being released as a completely separate platform.
OpenAI has explored generative music in the past, though those early projects came long before ChatGPT was launched. Lately, the company has focused more on audio technologies centered on text-to-speech and speech-to-text models.
Meanwhile, companies like Google, ElevenLabs, and Suno have also developed their own generative music tools. It's also worth noting that AI-generated content is already flooding streaming services, and let’s not forget the fiasco that was The Velvet Sundown. Unfortunately, this looks like just the start.
