According to The Information, OpenAI is actively working on its own neural network for music creation , which will compete with projects like Suno and Udio.
The publication’s sources report that the new product is expected to have several modes . First, it will simply generate a track based on a text request. Second, the model will be able to create music for a user-recorded voice track or video. This significantly expands the range of AI applications.
OpenAI is very cautious about copyright issues, especially after the launch of Sora 2, which initially generated copies of many famous characters, both real and fictional, from Stephen Hawking to Pikachu. To create the musical model, the company is collaborating with the Juilliard School, one of the most prestigious conservatories in the United States.
The school’s students help annotate the musical material: scores, structure, and harmonies . The idea is to train the model not simply from downloaded tracks, but using carefully annotated musical logic, which reduces legal risks and improves musical consistency.
It’s not yet known whether the new product will be released as a standalone app or integrated into one of OpenAI’s products. A logical candidate for integration would be Sora 2, which could overlay AI soundtracks onto videos created with it. The launch date for the new model has not been disclosed.
While OpenAI has launched generative music models in the past, these predate ChatGPT’s launch; more recently, the company has developed audio models focused on speech synthesis and text-to-speech.
Follow us on Google News to receive daily updates on cybersecurity. Contact us if you would like to report news, insights or content for publication.
