The British campaigners who helped defeat a plan to let AI companies train on copyrighted work by default are now warning Australia against adopting a similar rule, The Next Web reported Sunday, drawing on reporting by the Guardian’s Dan Milmo. Australia’s government is weighing an opt-out model that would allow AI firms to use creators’ work unless the creators object.

What Happened

The Australian proposal came to light this month through leaked slides from the Attorney-General’s Department, first reported by ABC News on Sept. 15. The slides described an opt-out approach, the model AI developers generally prefer because it places the burden of refusing on rights holders.

Beeban Kidron, the film director and House of Lords member who led the political fight against the UK version, directed her warning at Prime Minister Anthony Albanese, saying he should not waste his reputation by "selling Australian creatives down the river." Composer and former AI executive Ed Newton-Rex, who organized a silent protest album backed by more than 1,000 musicians and a wordless book signed by 10,000 writers, predicted that the public would quickly see the plan as unfair and said people expect AI companies to pay for the material they use.

RelatedNew York City Comptroller Urges Anthropic to Delay IPO Over Safety and Governance Concerns

Not everyone in the UK agrees with them. Antony Walker of the trade group techUK argued that rights holders had effectively shouted down the debate in Britain, and said AI training and fair payment for creators can coexist.

Attorney-General Michelle Rowland’s office said it is still consulting on options that would give creators meaningful control and fair payment. No decision date has been given.

Why It Matters

The UK reversal has become the main precedent against opt-out regimes. Britain’s government had favored the approach, but a consultation drew more than 10,000 responses and only 3% supported it. In March 2026, Technology Secretary Liz Kendall said the government no longer had a preferred option.

Australia is under direct pressure from the largest AI developers. OpenAI and Anthropic have both asked the government to ease copyright rules, and OpenAI has said it could not set up an Australian training centre without changes to the law. That gives the government an economic argument for moving toward an opt-out, and gives campaigners a specific commitment to challenge.

Context

Rules on training AI with copyrighted material are being settled country by country rather than through any international agreement. The choice between opt-in, opt-out and licensing-based systems determines who carries the cost of enforcement: under an opt-out, individual writers, musicians and artists have to find and flag their work, while under opt-in or licensing schemes AI developers must secure permission first.

The UK campaign succeeded partly through visible cultural protest, such as the silent album and the unsigned book, rather than through legal argument alone. Kidron and Newton-Rex are now applying the same public-pressure approach to a government that has not yet published a formal proposal.

What to Watch

Australia’s consultation is ongoing. The key signals will be whether the Attorney-General’s Department publishes the opt-out model as a formal option, whether it pairs any exception with a payment mechanism, and whether OpenAI or Anthropic tie specific local investments to the outcome. Any shift in tone from Rowland’s office after the UK campaigners’ intervention would show whether the British precedent is carrying weight in Canberra.