Australia’s music industry is escalating its opposition to the use of local artists’ work in artificial intelligence (AI) training datasets without authorization, following revelations that millions of songs have been included in such data collections. The issue has gained renewed attention amid ongoing tech industry lobbying efforts seeking copyright exemptions that could legitimize this practice.

Darren Hayes, an Australian singer-songwriter known for his solo career and as the former lead vocalist of Savage Garden, expressed strong dismay after an investigative tool developed by The Atlantic’s AI Watchdog uncovered four extensive song datasets circulated within the AI development community. Hayes took to social media to condemn the unauthorized use of his music, which spans three decades, stating that his entire catalog had been “stolen” and exploited to train AI software designed to generate music.

The AI Watchdog tool allows artists to search for their names to determine whether their works appear in these datasets, shedding light on the widespread nature of the issue. Hayes called for “the strongest legal retaliation possible” in response to these revelations.

Supporting this concern, the Australasian Performing Right Association and the Australasian Mechanical Copyright Owners Society (APRA AMCOS) confirmed that some of the most prominent Australian and New Zealand musicians were implicated. Names cited include Midnight Oil, Cold Chisel, Sia, INXS, Tina Arena, Kylie Minogue, Split Enz, Crowded House, Nick Cave, Tame Impala, Lorde, Flume, and The Cat Empire. APRA AMCOS noted that these findings were preliminary and that the full scope of the situation was expected to be considerably larger.

Dean Ormston, CEO of APRA AMCOS, emphasized the timing of these disclosures as significant, especially given ongoing government consultations driven by AI companies seeking exemptions from copyright laws. Ormston stressed that the works in question represent the “life’s work” of songwriters and are not mere bargaining tools.

For affected artists, the dispute transcends legal boundaries to touch on deeper issues of creative labor, artistic integrity, and personal identity. Hayes acknowledged that while AI cannot replicate the emotional authenticity derived from human experience, the unauthorized use of his music felt like a violation of his creative contributions.

As the debate continues, the Australian music community appears poised to advocate for stronger protections and clearer regulations governing the use of copyrighted material in AI technologies.