thepublishingtimes.comEditorial

The Publishing Times

All the News That’s Fit to Publish

publishing-news · Guides

Anthropic Tells Australia's AI Inquiry That Licensing Everything It Trains On Is 'Technically Impossible'

At an Australian parliamentary AI hearing, Anthropic said current copyright law makes local model training impossible. Creative groups said the law should stay as it is.

The Publishing Times Staff · October 7, 2026 · 5 min read

Anthropic Tells Australia's AI Inquiry That Licensing Everything It Trains On Is 'Technically Impossible'

Should AI companies have to license the books they train on? That question dominated a hearing of the Australian Parliament's Joint Select Committee on Artificial Intelligence in Sydney on Tuesday, October 6, 2026 (Sydney time). There, Anthropic, the maker of the Claude chatbot, said Australia's current copyright law makes it impossible to train AI models in the country.

According to The Guardian, Anthropic's special envoy Jeffrey Bleich, a former U.S. ambassador to Australia, told the committee the company had "never tried to dictate to Australia" on copyright. "We've just shared information about how the training works, and then discussed what its implications would be with respect to our capacity to operate here," he said.

The Guardian reported that Anthropic's position is that current Australian law would require AI companies to obtain a licence for every piece of content on the internet used as training material, which Bleich called "technically impossible." The Sydney Morning Herald reported that, under questioning, Bleich said it was technically impossible to "license the entire internet," and that training in Australia could be found illegal under existing law without a licence from every rights holder. The Herald said he urged Australia to host the training of advanced AI models "under its own laws, with its own conditions attached."

The opt-out fight

The Guardian said the morning session focused largely on copyright licensing, "amid concerns companies such as Anthropic were pushing the Australian government to adopt an opt-out model," under which rights holders would have to opt out specifically to stop AI training on their work.

Creative and media groups rejected that approach:

  • The ABC told the committee Australia's existing copyright regime was "completely adequate," according to The Guardian. Kate Gilchrist, its head of content and legal operations, said an opt-out system would mean checking every site where ABC material appears. "We cannot scour the internet and ensure we are opting out on all of those different sites of publication. It simply does not work," she said.
  • The Australian Copyright Council's chief executive, Eileen Camilleri, said letting developers use creative work without permission or payment posed "an existential threat" to the sector, the Herald reported.
  • Aria, the recording industry body, said through its chief executive Annabelle Herd that "Australia's artists will be the roadkill in the rush" to make deals with AI companies, and that Anthropic and OpenAI were "asking the government for a free ride," per The Guardian.
  • The Media Entertainment and Arts Alliance called for an equitable remuneration scheme that would put fees on companies across the AI technology stack, with the money going to creative and media workers and to a cultural fund, The Guardian reported.

The Herald reported that groups representing about 128,000 Australian songwriters and more than 40,000 authors and artists gave evidence before lunch. It also reported that the committee at one point moved Anthropic's session behind closed doors, and that independent senator David Pocock, a committee member, said the frontier developers had provided "no cohesive argument" for why copyright changes were needed.

Where Australian policy stands

The government has already said no to the broadest version of what tech companies wanted. In an October 26, 2025 release, Attorney-General Michelle Rowland said the government's consultation on possible copyright updates "will not include a Text and Data Mining Exception," the kind of carve-out that would let AI developers use Australian creators' work "for free and without permission." The Herald reported that Camilleri welcomed that decision at the hearing.

The joint committee was appointed on August 20, 2026, and must report by November 30, 2026. Its hearings continue on October 7 in Sydney and October 8 and 9 in East Melbourne. A transcript of the October 6 hearing is listed on the committee's public hearings page.

Why it matters for authors

Licensing versus opt-out is the live question. Anthropic's envoy says licensing every work used in training is impossible. Creative groups say the current law already works, and that an opt-out system would shift the job of protecting a book onto its author or publisher.

Australia is a test case. The same licensing-versus-opt-out argument is running in other countries, so authors outside Australia have reason to watch what the committee recommends in November.

Know your rights position now. Check your publishing contracts for AI training clauses, and watch for guidance from the Australian Society of Authors, the Authors Guild, and the Society of Authors if you publish in those markets.

Sources