Australian musicians including Paul Dempsey and Bernard Fanning discovered their original songs were present in datasets used for training artificial intelligence between August and October 2025. Songwriter Darren Hayes found the entire output of his 30-year recording career in these datasets.

The discoveries were made possible through a dataset search tool that identifies creative works scraped from the internet for AI training. Dempsey located the complete catalogue of his band Something For Kate and his solo tunes using this tool. Hayes' recording career includes Savage Garden hits such as "Truly Madly Deeply."

Dempsey stated, "It's frustrating this is happening. Every negotiated agreement and contract I've ever gone into in my career with whatever entity or record label is all just rendered useless." He added, "An artist's ability to negotiate fair terms for the use of their content is just being ripped away from them." Regarding the future of art, Dempsey said, "We can trigger huge emotional responses in each other through art, and I don't know that that's going anywhere; it's just going to be flooded with all this other shit."

Hayes stated, "I absolutely feel violated that all of the hundreds and hundreds and hundreds of hours, blood, sweat and tears that I've put into my music, along with every other musician, has been stolen and served up like french fries to a piece of software that spits out shit."

Dean Ormston, chief executive of APRA AMCOS, an organization representing 128,000 members in Australasia, said, "Major tech platforms have not come to the table. Not once. Instead, they have lobbied governments, circulated policy papers, and proposed solutions designed to extinguish any obligation to pay."

The Australian songs were found within two datasets, Sleeping-DISCO-9M and LAION-DISCO-12M, assembled by researchers including Sleeping AI and the Germany-based group LAION. The Sleeping-DISCO-9M dataset contains 9.7 million music tracks from YouTube and lyrics from Genius.com. The LAION-DISCO-12M dataset includes 12.3 million YouTube tracks. Inclusion in datasets is not definitive proof of use, as AI companies may omit works when training their models.

Australia's intellectual property laws mandate permission and agreed-upon terms, including payment, for the use of copyrighted works. The IT industry has advocated for exemptions to these laws concerning text and data mining. In August 2025, the Productivity Commission suggested changes that would have legalized AI companies using content without compensating creators. The Australian federal government subsequently rejected these proposed changes in October 2025.