AI Watchdog stems from an investigation from The Atlantic into AI data training.
The Atlantic has unveiled a new tool that allows artists to check whether their works are present in data sets used for AI training.
Compiled by researcher Alex Reisner, AI Watchdog draws on four giant datasets spanning millions of songs that are being shared within the AI-development community. Tracks from millions of artists in the electronic and dance music community can be found via the tool, from its superstars to its underground and independent artists.
AI Watchdog was established in 2025 to highlight the millions of books, research articles and video media that has been used to train AI products. It’s now been expanded to include music datasets. The tool isn’t comprehensive. It only draws upon four datasets that Reisner found in research papers and AI data-sharing sites. One dataset contains 12 million tracks, while another contains 9 million tracks. The two other datasets both contain over 100,000 tracks.
Companies like Google and Stability have openly utilised these datasets, but due to the AI industry’s opaque mechanics, it’s currently unclear who has used these datasets, and how many others like it may exist. A track appearing in the tool doesn’t definitively mean it was used to train AI, but due to the existence of other datasets, a track’s absence is also not evidence it hasn’t been used.
In three of the datasets, songs were distributed via links to YouTube and Spotify, which are often accessed using “tools that automate the job, some of which allow developers to bypass logins, advertisements and mechanisms that might earn money or subscribers for creators,” according to Reisner. A fourth dataset draws upon the Free Music Archive collection, a project initially founded by New Jersey radio station WFMU.
Typically, the training data that is fed into generative-AI models are closely guarded secrets. Many AI companies have claimed that the data, in this case recordings, are proprietary. Google, for example, trained its audio-generating models on “materials that [Google subsidiary] YouTube and Google has a right to use under our terms of service,” according to a blog post from the company.
Some major labels have sued AI music companies like Suno and Udio. Others, like Warner Music Group and Universal Music Group, have settled or established licensing and partnership agreements with the companies. Streaming services have slowly begun tagging AI-generated content on their platforms, including Deezer, who recently launched an AI music detector for all streaming platforms.
Check out the tool, and browse Instagram for more information.
https://www.theatlantic.com/technology/2026/06/dataset-free-music-archive/687336/
Leave A Comment
You must be logged in to post a comment.