The Atlantic has published four searchable databases exposing the massive scale of AI music training. One database alone contains 12 million tracks, while others hold 9 million and roughly 100,000 songs each. These catalogs reveal copyrighted music feeding machine learning models without explicit permission.

Staff writer Alex Reisner found hits from major artists like Taylor Swift and Bad Bunny among the scraped content. Legal battles already target generative AI platforms such as Suno and Udio, which claim fair use protections. However, a recent book publishing case settled for $1.5 billion, suggesting stronger copyright arguments may succeed.

Streaming services have attempted to flag AI-generated content with mixed results, yet imposter bands still profit from machine-made knockoffs. The databases provide concrete evidence that music industry lawyers could leverage for similar litigation, potentially reshaping how AI companies access copyrighted creative works.