Australian musicians sound warning note after Nick Cave, Kylie and many more slurped into AI training tool
TL;DR
The Atlantic has built a search tool for AI training datasets. According to the report, it surfaces songs by Nick Cave, Kylie Minogue, Powderfinger, Jimmy Barnes, Bernard Fanning, Paul Dempsey and Darren Hayes. The works appear in two music datasets: Sleeping-DISCO-9M, with 9.7m YouTube tracks plus Genius lyrics, and LAION-DISCO-12M, with 12.3m YouTube tracks.
Nauti's Take
This is not a minor culture-war spat between nostalgic musicians and new tech. When AI companies ingest training data first and talk about exemptions, fair use or opt-outs later, the market is already distorted.
The PR line about a creative revolution sounds thin when the raw material is collected without clear consent. Innovation needs data, but data rights are not paperwork to be brushed aside.
Briefingshow
Music is not just data; it is a market built on rights, identity and recognizable creative choices. If AI systems train on millions of songs without artists knowing whether or how their work was used, the power balance shifts: platforms scale first, creators negotiate later. This is where AI music either becomes a tool for artists or an extraction model built on them.
Sources