The platform now allows artists to check whether their recordings appear in datasets used within the AI development community.
Photo credit: The Atlantic – Website
The Atlantic has expanded its AI Watchdog project with a new music-focused tool that enables artists to search whether their recordings appear in datasets commonly referenced in AI research and development.
Compiled by researcher Alex Reisner, the database draws on four large music datasets containing millions of tracks. The tool includes recordings from major artists, independent musicians, and underground producers across electronic music and beyond.
Originally launched in 2025 to highlight books, research papers, and video content used in AI training, AI Watchdog now extends its scope to music. The datasets referenced include collections containing up to 12 million and 9 million tracks, respectively, alongside two smaller archives with more than 100,000 recordings each.
The project does not confirm whether a specific track was used to train an AI model. Likewise, the absence of a recording from the database does not guarantee that it has not been used elsewhere. Reisner notes that several datasets were assembled using links to platforms such as YouTube and Spotify, while another draws from the Free Music Archive collection.
Check out AI Watchdog here.
Ver esta publicación en Instagram
