Atlantic Reporter Publishes Searchable Lists of Four Music Datasets Used in AI Research

Atlantic reporter Alex Reisner uncovered four datasets of music used to train AI models and published them in a searchable format on the Atlantic’s AI Watchdog site. Two datasets contain 12 million and 9 million tracks respectively, while the remaining two each hold more than 100,000 songs. The Verge reported that the four datasets have been downloaded thousands of times.

Google confirmed it has used the datasets in research papers. The Verge reported that three of the datasets are distributed as lists of links to songs on YouTube or Spotify.

The Free Music Archive dataset is one of the four. The Verge reported that some sources in the datasets, such as the Free Music Archive, allow free streaming for personal use but require licensing for commercial applications. AI developers download the actual audio using automated tools that can bypass logins, advertisements, and payment mechanisms, which violates the terms of service of these platforms.

Radiohead, Aphex Twin, Wu-Tang Clan, Bruce Springsteen, and experimental composer Hainbach. You can search the songs, books, and other media on the Atlantic’s AI Watchdog site.