The Atlantic Created a Searchable Database of the Music Used to Train AI

The Atlantic released a searchable database of music used to train AI models. Reporter Alex Reisner uncovered four datasets used to train AI systems. Two of the datasets are enormous—12 million and 9 million tracks. The other two are much smaller but still represent a significant amount of training data: one contains over 11,000 tracks and the other over 4,000 tracks. The entire collection has been made fully searchable for the public.

This database matters because it reveals the scale and diversity of music feeding into AI models. For artists and music lovers, it's a way to see what songs and genres are shaping AI's understanding of music. It also raises questions about copyright and compensation for artists whose work is used in training without permission.

You can explore the database today by visiting The Atlantic's dedicated page. Simply search for artists, albums, or genres to see what tracks are included in these training sets. It's a fascinating way to understand the musical DNA of AI.