An award win is usually a moment to celebrate. For the Australian creative known professionally as Partos, it has instead become a platform to reopen one of the most contentious debates in the local technology and arts worlds: who gets to decide how artificial intelligence systems feed on copyrighted work, and who gets paid when they do.
According to IT Brief Australia, Partos used the recognition to argue that Australia can no longer afford to treat AI and copyright as a problem for another day. The message, in essence, is that the country’s creators are watching their work become raw material for machines that were trained without their consent, and that the policy settings have not kept pace.
Why this keeps surfacing
The timing is not accidental. Over the past two years, generative AI tools have moved from novelty to everyday utility, and nearly all of them were built on vast quantities of text, images, music and video scraped from the open web. Much of that material is protected by copyright, and much of it was created by people who never agreed to hand it over for training. That tension sits at the heart of court cases, licensing fights and policy reviews across the United States, Britain, the European Union and, increasingly, Australia.
Locally, the question has been simmering inside government for some time. The Attorney-General’s Department convened a Copyright and Artificial Intelligence Reference Group to work through exactly these issues, bringing creators, publishers, platforms and legal experts to the same table. Separately, the Productivity Commission waded into the debate during its work on harnessing digital technology, floating the idea of a text and data mining exception that would make it easier for AI developers to use copyrighted material for training. That suggestion landed badly with the creative sector, which saw it as handing away the value of Australian work to overseas technology companies.
Against that backdrop, a respected award winner stepping up to demand action is more than a personal opinion. It is a signal that frustration inside the creative community has not eased, and that individual practitioners are no longer content to leave the argument to industry bodies and lawyers.
The news
What makes the intervention notable is who is making it and when. Awards nights are shop windows for an industry, moments when the sector presents its best work to clients, peers and the public. Choosing that stage to talk about copyright and AI is a deliberate decision to put a commercial and ethical problem in front of an audience that would rather be toasting the winners.
The thrust of Partos’s argument, as reported, is that Australia needs clearer and firmer rules on how AI systems may use creative work, rather than a loose arrangement that quietly favours the companies building the models. That lines up with a broader push from Australian creators for three things: transparency about what has been used to train commercial AI products, consent before protected work is ingested, and compensation when it is. Those principles have become the rallying points for musicians, writers, illustrators, photographers and screen workers who feel the ground shifting beneath them.
Two ways of seeing it
From the creative side, the case is straightforward. A song, a photograph or a piece of writing represents years of skill and investment, and if a machine can absorb thousands of those works and then generate convincing imitations, the market for the originals is undercut. Rightsholder groups argue that existing copyright law already requires permission to copy protected material, and that training a model on that material without a licence is not a loophole but an infringement dressed up as innovation. The music sector has been especially vocal, having put a figure on what is at stake for creator revenue as AI-generated tracks proliferate.
The technology industry sees a different picture. Developers and many economists argue that restrictive copyright rules will simply push AI research and the jobs that come with it offshore, to countries with more permissive regimes. They point to the practical impossibility of licensing every item in a training set the size of the internet, and warn that Australia risks becoming a consumer of foreign AI rather than a builder of its own if it clamps down too hard. A text and data mining exception, on this view, is not a giveaway but a competitiveness measure that keeps the country in the game.
Government is caught in the middle, trying to protect a creative economy worth tens of billions of dollars while also encouraging an AI sector it has publicly identified as central to future productivity. There is no settled position yet, which is precisely why voices like Partos’s carry weight. Each high-profile intervention nudges the debate and reminds policymakers that whatever they decide will have winners and losers with real names.
The Australian stakes
For Australia, this is not an abstract overseas quarrel. The country’s creative industries employ hundreds of thousands of people and feed into advertising, film, music, publishing, gaming and design, sectors that already operate in the long shadow of much larger international markets. If the rules allow unlicensed training by default, Australian creators could find their work powering tools owned and monetised elsewhere, with little flowing back home.
There is also a sovereignty dimension that policymakers have begun to take seriously. Much of the most capable AI is built by a handful of global firms, and Australia has limited leverage over how those companies behave. Domestic copyright law is one of the few levers genuinely within the country’s control, which is why creators are so determined that it not be weakened. The alternative, they argue, is a future in which Australian culture is freely harvested and the economic benefit accrues to Silicon Valley rather than to the studios of Surry Hills, Fitzroy or Fortitude Valley.
For businesses deploying AI, the uncertainty is its own problem. Agencies and companies building products on top of generative tools want to know whether the content those tools produce is legally clean, and whether they could be exposed if the underlying model was trained on infringing material. Clearer rules would help them as much as they would help the creators who make the original work.
What’s next
The practical question is whether political attention translates into legislation or merely more consultation. The Copyright and AI Reference Group’s work continues, and any move toward a text and data mining exception would almost certainly trigger a fierce public fight given how strongly the creative sector has already reacted. In the meantime, expect more licensing deals struck directly between rightsholders and AI developers, more individual creators speaking out as Partos has, and growing pressure for transparency requirements that would force companies to disclose what their models were trained on.
None of that resolves the core tension between protecting creators and fostering an AI industry, but it does keep the issue where campaigners want it, firmly on the national agenda rather than filed away for a quieter year. An award speech will not change the law on its own. It is, however, a reminder that the people whose work is at stake intend to keep making noise until Canberra decides whose interests the rules will serve.
Sources: IT Brief Australia.
