Overview
- The International Confederation of Music Publishers released findings alleging major AI firms trained models on copyrighted songs, lyrics and vocal styles without permission.
- Investigators say material was scraped by web-crawlers from platforms such as YouTube and other licensed services, with no licenses secured or royalties paid.
- Publishers argue the practice enables AI outputs that closely imitate famous artists, characterizing it as commercial exploitation of creative work.
- Legal and commercial responses are advancing, including the RIAA’s 2024 lawsuit against Suno and Udio, additional actions by major labels, and ongoing talks to establish licensing frameworks.
- A $1.5 billion Anthropic settlement over scraped books is cited as a precedent, while ICMP warns of a potential 20% income hit for creators and points to the EU AI Act as a model for transparency requirements.