Music datasets used to train AI models are now freely available for public search. The datasets, created by Atlantic reporter Alex Reisner, contain millions of tracks from popular artists like Lady Gaga and Radiohead. Three of the datasets are distributed as links to songs on YouTube or Spotify, which can be downloaded using automated tools that bypass platform rules. This raises concerns about copyright infringement and the use of copyrighted materials in AI development. Why it matters: The widespread use of these datasets highlights the need for greater transparency and regulation in the use of AI training data.