OpenAI hunts for music’s last human note
OpenAI plans to develop a generative artificial intelligence music tool that creates soundtracks from text or audio prompts, according to The Information. The company behind ChatGPT would compete with platforms such as Suno and Udio, which face lawsuits from major record labels over alleged copyright violations in their training methods.
The proposed tool could add music to videos or instrumental backing to vocal recordings. Students at the Juilliard School have assisted OpenAI by annotating musical scores for training data, though the company has not disclosed its sources of music content or a release date for the product.
OpenAI faces multiple copyright infringement lawsuits from publishers and licensing groups for its unauthorized use of protected material. The company reached a licensing agreement with Shutterstock that provides access to audio files from the stock media library.
Artificial intelligence music floods streaming platforms at an accelerating rate, with Deezer reporting that 28 percent of uploaded tracks in September were fully computer-generated, up from 18 percent in April.
