For the better part of a year, Australia’s creative community has been shouting into a policy vacuum about artificial intelligence, and the chorus is starting to sound a lot like a John Farnham anthem. Now a group of Australian academics has offered something more structured than protest: a proposed plan aimed at protecting local artists from having their work, voices and likenesses swept up and exploited by generative AI systems.
The reporting, published by Cyber Daily under the cheeky banner “You’re the Voice,” lands at a moment when the tension between the technology sector and the creative industries has rarely been higher. Generative models are trained on vast troves of text, images, music and video, much of it scraped from the open internet without the knowledge or consent of the people who made it. For working musicians, writers, illustrators and performers, that practice feels less like inspiration and more like uncompensated extraction, and the legal tools to do anything about it have not kept pace.
Why the timing matters
This is not an isolated flare-up. FluentSea has tracked a steady drumbeat of creative-sector pushback throughout the year, from the ARIA charts moving to exclude fully AI-generated songs, to screen composers protesting in defence of human-made music, to Spotify pulling AI acts from its playlists. A recent poll even found that a clear majority of Australians side with creators in the copyright fight. The academic proposal is significant because it tries to translate that broad public sympathy into a workable set of rules rather than a slogan.
At the heart of these proposals, as they have surfaced in Australian academic and legal circles, sit a handful of familiar principles. Artists and rights holders want consent before their work is used to train a model, credit so audiences can tell human output from synthetic, and compensation when their creative labour powers a commercial product. The sticking point has always been enforcement. Copyright law was built for an era of copying and distribution, not for machine learning that absorbs millions of works and emits something statistically derived from all of them at once.
Two sides of a hard argument
The creative industries argue the status quo is untenable. Without a clear right to say no, and a mechanism to be paid when the answer is yes, they say the local arts economy risks being hollowed out by systems that learned their craft for free. Bodies representing musicians and performers have warned for months that voice cloning and style mimicry pose an existential threat to the value of an artist’s identity, which for many performers is the only asset they truly own.
The technology camp pushes back with equal conviction. Developers and some economists contend that overly tight restrictions, particularly a strict opt-in regime for training data, would strand Australian AI builders behind the United States and parts of Asia, where fair-use style doctrines give companies far more room to move. The Productivity Commission has floated the idea of a text and data mining exception precisely to keep local firms competitive, a suggestion that drew a furious response from artists who saw it as legalising the very scraping they want stopped. That clash frames the academic plan: it is an attempt to find a path that protects creators without smothering the domestic AI sector in the cradle.
What the plan is reaching for
Proposals of this kind typically call for a mix of reforms rather than a single silver bullet. They tend to include a specific right for performers to control the use of their voice and image, transparency obligations that would force AI companies to disclose what they trained on, and some form of collective licensing so that payment can flow at scale rather than artist by artist. The attraction of a collective model is pragmatic. Individual creators cannot realistically negotiate with trillion-parameter systems, but a licensing body can, much as organisations already collect and distribute royalties for radio play and public performance.
The challenge is that none of this works unless Canberra writes it into law. Academic frameworks can shape the debate and give government a ready-made blueprint, but they carry no force on their own. That puts the ball squarely in the court of the federal Attorney-General and the government’s copyright and AI reference group, which has been weighing exactly these questions.
What it means for Australia
For a mid-sized economy with an outsized cultural footprint, the stakes are real. Australian music, film and design punch well above their weight globally, and the sector employs hundreds of thousands of people, many of them freelancers with little bargaining power. If local artists cannot protect their work from being ingested and regurgitated by AI, the economic incentive to keep creating in Australia weakens, and the distinctive voice of Australian storytelling could be diluted into a global average generated offshore.
There is also a sovereignty angle that fits the broader national conversation about controlling the terms of AI adoption. A homegrown framework, designed around Australian values and Australian law, would let the country set its own balance between innovation and protection rather than importing rules written for Silicon Valley. The counter-risk is that if the settings are too protective, AI development simply happens elsewhere and Australian creators end up competing with imported synthetic content anyway, only without any of the guardrails.
What’s next
The immediate value of the academic proposal is that it gives policymakers a concrete starting point as consultations continue. Expect the usual players, the music and screen peak bodies, the tech industry associations and the legal academy, to line up with submissions, and expect voice and likeness protection to be the flashpoint given how directly it touches performers. Whether any of it becomes legislation before the next election is an open question, but the direction of travel is clear. The days of treating creative work as free training fuel are ending, and the only real debate left is how quickly, and on whose terms, that reckoning arrives.
Sources: Cyber Daily.
Comments 1