OpenAI is reportedly developing a new artificial intelligence tool capable of generating music from text and audio prompts. According to The Information, the company is said to be collaborating with The Juilliard School in New York to build a large database for training the system. Students are reportedly helping by annotating musical scores to improve the model’s understanding of music.
However, Juilliard has publicly denied any partnership, telling Euronews Next that the reports are “not true.” OpenAI has yet to respond to requests for comment.
This would not be the company’s first step into AI music. In 2019, it introduced MuseNet, a neural network that composed four-minute pieces with up to ten instruments. A later tool, Jukebox, released in 2020, could produce songs with basic vocals across several musical genres.
The latest development arrives amid growing industry tension. Other major technology firms, including Spotify, are racing to launch their own AI music tools. Spotify recently partnered with Sony, Universal, and Warner to develop “responsible AI products,” while also integrating AI into playlist creation and its DJ feature.
Yet, artists and music groups are voicing strong copyright concerns. European organisations, including the European Composer and Songwriter Alliance (ECSA) and GESAC, argue that the EU AI Act fails to protect creators.
Several lawsuits are already challenging AI companies. The Recording Industry Association of America (RIAA) is suing US start-up Suno, accusing it of using pirated songs. Universal Music Group has also filed a copyright lawsuit against AI firm Anthropic, warning that creators’ rights are increasingly at risk.
