What happens when an AI-generated artist gets a record deal? A copyright mess (www.theverge.com)

A small record deal has exposed a large legal gap: Hallwood Media signed Telisha “Nikki” Jones — credited as the lyricist for the AI-generated R&B act Xania Monet — even though the avatar, vocals and full songs appear to have been produced by the AI music tool Suno. Industry scrutiny centers on what, exactly, the label bought: copyright law currently recognizes only human-authored elements, so Jones’s lyrics may be protectable but the AI-generated composition, performance and sound recording likely are not. Suno has been sued by major labels and admitted training its model on music scraped from the web, heightening infringement concerns. The case matters because it highlights practical and technical implications for artists, labels and lawmakers. Without clear copyright in AI outputs, buyers risk paying for assets they can’t enforce or exclusively license, and platform credits or streaming metadata don’t substitute for registration. The US Copyright Office’s guidance rejects authorship from mere prompting, and Congress is debating access to AI training records. More broadly, this deal underscores a cultural shift: generative audio models can produce commercially viable music, but unsettled IP rules, training-data provenance and enforcement gaps will reshape how the music industry values, licenses and litigates AI-created work.