Australian Musicians Slam AI Training Tools for Using Their Songs Without Permission

A growing number of Australian musicians are speaking out against the use of their original songs in datasets to train artificial intelligence. Paul Dempsey and Bernard Fanning are among the big-name artists who have discovered their music in these datasets, which were uncovered by a search tool created by The Atlantic. The datasets, compiled by researchers and a German-based group, contain millions of creative works, including songs by Kylie Minogue, Powderfinger, and Jimmy Barnes, as well as novels by Thomas Keneally and Peter Carey. Dempsey expressed frustration that his entire catalogue, including his work with Something For Kate and solo tunes, was found in the datasets without his permission.

He believes this practice, known as 'slurping', undermines an artist's ability to negotiate fair terms for their work. Bernard Fanning argued that using original songs to produce AI content is dehumanizing, as it replaces human emotion and experience with robotic aggregation. Other musicians, such as Darren Hayes, have also spoken out against the practice, feeling violated that their life's work has been used without permission. The issue has sparked concern about copyright and intellectual property laws, with music licensing organization APRA AMCOS stating that major tech platforms have not taken responsibility for paying creators for their work.

Australia's intellectual property laws require permission and agreed-upon terms before copyright works can be used, but the IT industry has pushed for exemptions. The federal government recently ruled out changes that would have legalized AI companies using content without paying creators. Musicians like Dempsey argue that genuine artistic expression comes from human experience, not artificial intelligence, and that the use of their songs without permission is a violation of their rights.

Source: The Guardian