Copyright Plus Gives Creators Power Over AI Use

A new framework called Copyright Plus would give Australian creators the right to refuse AI training use of their work, share in the value it creates, and seek fast removal of imitations. Academics at UNSW Sydney have presented the proposal to the federal government as a bolder answer to a problem artists say has run ahead of the law: putting work online now means losing control over what happens to it next.

The proposal arrives as AI companies themselves begin to acknowledge the disruption their technology causes. Anthropic recently announced a public benefit fund for research into labour market disruption, while conceding that bolder approaches are needed given the pace of AI development. The framework is pitched as exactly that kind of approach, one that keeps AI development going without treating creative work as free raw material.

Three powers for creators

At its core, the framework would hand creators three kinds of power they largely lack today. The first is the right to say no. Under the proposal, AI companies would have to disclose their data sources, and rights holders would have to tell original creators when their work has been licensed for AI use, including the payment share and any terms affecting their moral rights. "Artists simply don't have power over how technology companies use their work, a meaningful say in how media companies license copyright in their works, nor any power to remove unauthorised synthetic works from circulation," said Professor Kathy Bowrey, an intellectual property specialist at UNSW Law.

The disclosure duty matters because copyright in a creative work is not always held by the person who made it. A publisher, record label or production company may hold the rights, and under current Australian law, AI companies do not have to disclose data sources, while rights holders have no duty to inform original creators about AI licensing. Copyright Plus would close both gaps at once, and every company involved in licensing would have to honour a creator's refusal to have work scraped for training or mimicked in synthetic output.

The second power is payment. The framework would impose a levy on AI companies to fund fair payment for creative professionals whose work was likely used in the past without permission, alongside a broader cultural fund supporting the arts ecosystem. Unions and professional organisations would distribute income from a fair remuneration fund, reaching people who may never land a major commercial deal, including performers, voice artists, session musicians and visual artists. A relatively unknown visual artist, for example, could still receive payment through their professional organisation even without ever signing a big licensing agreement.

Enforcement that artists can actually use

The third power is enforcement, and the researchers argue it may matter most. A right offers little protection if enforcing it requires time and money an artist does not have. Copyright Plus therefore creates faster ways for artists and Indigenous creators to flag unauthorised AI content and have it removed, including works that mimic real creators. The key device is an expedited presumptive takedown notice, which would shift the burden of proof onto the party distributing the content to show it is not ripping off authentic work.

In practice, artists could notify their professional organisation about fake websites, synthetic social media posts that mirror authentic artists, or content licensed under a "No AI derivative" condition or subject to Indigenous cultural and intellectual property restrictions. After basic verification, the organisation could ask the host to remove the offending material, in the same way media companies already issue takedowns for copyright infringement. The framework would be funded by AI corporations through an AI Operations and Licence Fee, so the cost of policing the system would not fall on individual creators.

Supporters point out that mandatory reporting of training data sources is already part of law in the European Union and California, and that overseas copyright litigation is exposing unauthorised use of Australian content. The proposal does not try to turn back the clock on AI, its authors stress: it accepts that the technology is here, and instead asks who benefits when machines learn from human creativity. The counter-argument, familiar from every copyright fight, is that heavier duties could slow innovation or push development elsewhere, which is why the framework pairs new obligations with a levy-funded system rather than outright bans.

Why Australia, and why now

The proposal draws an explicit parallel with an earlier technological upheaval. Almost a century ago, radio gave rise to public broadcasting, and the ABC was built with a charter to serve national identity, cultural diversity, education and the arts. The lesson, according to Associate Professor Oliver Bown, a co-author of the proposal, is that technologies built with public money should serve the public good. "If we don't design people power into our AI policies, we resign ourselves to being dictated to by foreign corporate powers," Bown said.

For readers, the stakes are closer than they look. The songs, images, stories and performances that fill feeds and playlists are the training material of the next generation of AI tools, and the rules now being written will decide whether the people who made them get a say, a share, or nothing at all. The framework is only a proposal, not law, and its fate now rests with the government it was presented to. But it gives the debate something it has lacked: a concrete design, with named duties, a funding mechanism and a fast enforcement path, rather than another round of warnings about a future that has already arrived.

The debate lands as AI tools reshape everyday work, from an intelligent assistant arriving in Apple's office software to workday chores quietly reshaping the workweek. Those shifts show why the question of who controls creative work matters beyond studios and stages: when machines can imitate, summarise and reproduce human output, the rules governing that output become rules about livelihoods.