Anthropic is pushing Australia to clarify intellectual property rules to enable training of its AI models and to support data-centre investments in the country, as it signals a conditional commitment hinged on a clearer copyright framework.

Australia’s Labour government faces pressure from musicians, screenwriters, and artists to reject proposals that would let AI models train on copyrighted works for free, adding political weight to the reform debate.

Dario Amodei met with Treasurer Jim Chalmers in April to discuss entering the Australian market and potential data-centre infrastructure, framing investment as contingent on stable copyright policy.

The Attorney-General’s department reportedly ruled out a text-and-data mining exception, shaping the reform pathway and enforcement considerations for rights holders.

Canberra has signaled it is engaging with multiple stakeholders on the IP issue, while Anthropic has not yet commented publicly.

Industry players and rights holders warn against reopening copyright settings, arguing current licensing arrangements with AI developers work across news, music, and publishing sectors.

Official briefings indicate Australia will not adopt a text-and-data mining exception and that Anthropic is actively engaging with stakeholders on licensing structures.

Officials say Anthropic is in talks with various stakeholders and would proceed only with clear copyright rules and viable licensing pathways.

Australia’s position remains that a TDM exception will not be introduced, with Anthropic continuing engagement on licensing amid policy uncertainty.

Officials note Anthropic’s concern about a “long tail” of smaller rights holders complicating licensing and data procurement.

The briefing emphasizes that the ability to license data and rights, including the small-rights challenge, will influence Anthropic’s approach in Australia.

Anthropic recognizes licensing hurdles tied to a broad range of smaller rights holders and the difficulty of identifying eligible licenses in Australia.