The music industry's worst fears have been realised, with millions of songs, including chart-topping hits from Australian artists like Kylie Minogue and Nick Cave, being scraped into datasets used to train artificial intelligence models. A search tool developed by The Atlantic has uncovered the extent of this data harvesting, revealing that 9.7 million music tracks from YouTube and 12.3 million more from LAION's dataset have been aggregated for AI training purposes. This raises fundamental questions about copyright, fair compensation, and the very essence of artistic expression in an age where machines can mimic human creativity.
Australian musicians are taking a stand against this practice, with Something For Kate's Paul Dempsey and Powderfinger's Bernard Fanning among those speaking out against what they see as the dehumanisation of art. 'It's like our entire careers have been ripped away from us,' Dempsey said, highlighting the inadequacy of existing contractual agreements to safeguard artists' rights in the face of AI-driven content creation. Fanning concurred, arguing that robots lack the capacity for genuine artistic expression and that human feelings are what make art meaningful.
The datasets in question – Sleeping-DISCO-9M and LAION-DISCO-12M – contain a significant portion of Australia's musical heritage, including decades' worth of work from prominent artists. APRA AMCOS, which represents 128,000 members in Australasia, considers the scraping of creative works as 'theft', with chief executive Dean Ormston lamenting major tech platforms' failure to engage in meaningful negotiations over usage and compensation.
While Australia's intellectual property laws generally require permission for copyrighted work use, industry exemptions for text and data mining have been proposed but rejected by the federal government. This decision follows a report from the Productivity Commission in August 2025 advocating for AI companies' ability to use content without paying creators. The repercussions of this incident extend far beyond Australian shores, posing an existential threat to creative industries worldwide.
For UK businesses and consumers, the implications are stark: a potential deluge of AI-generated music could dilute the market for human-created art, making it increasingly difficult to distinguish between the two. Economically, if creators are not compensated fairly, the value of original content will be eroded, with long-term consequences for the creative industries' contribution to UK GDP.