Albanese and Taylor edge toward AI copyright deal as One Nation rises

Bipartisan plan would compensate creatives without forcing OpenAI and Anthropic into individual licences, as Greens and Senator Pocock warn on payouts

By LineZotpaper
Published
Read Time2 min
Prime Minister Anthony Albanese and Opposition Leader Angus Taylor are moving closer to a bipartisan deal that would make Australia more attractive for the world's largest artificial intelligence companies, amid warnings from Anthropic that copyright laws are blocking investment and criticism from the Greens and Senator David Pocock.

Albanese and Taylor met at Parliament House on Monday, according to sources in both parties, as the leaders narrowed differences over how creative rights holders should be compensated for content used to train AI models such as Anthropic's Claude.

The two leaders have been meeting more regularly than Albanese did with former Liberal leader Sussan Ley, and both sides are eager to settle a policy that compensates rights holders without forcing OpenAI and Anthropic to sign individual licence agreements. Sources in both parties said such an outcome would create an unwanted global precedent and could lead to international rights holders suing AI firms under Australian law.

Labor has not yet put a clear proposal to the Coalition, so there is no final agreement. A dispute is possible once the government settles on a position.

The Greens and Senator Pocock have raised concern that a Labor-Coalition deal would not properly compensate creatives whose work is used to train models such as Claude.

The negotiations come as One Nation continues to out-poll the Coalition on first preferences and sits just four percentage points shy of Labor. Albanese lamented on Tuesday that Australian politicians were appealing 'to the fringes'.

The Age reported in August that Coalition home affairs spokesman Jonno Duniam had extended an olive branch to Labor on AI regulation in a bid to ease public anxiety, and Anthropic has reasserted its warning that Australia's copyright laws are blocking investment.

§

Analysis

Why This Matters

  • Australia's copyright rules will help decide whether AI companies such as OpenAI and Anthropic build major operations locally rather than in other markets.
  • A bipartisan deal could set a precedent for how countries compensate creatives for AI training data without scaring off investment.
  • The negotiations are unfolding under electoral pressure, with One Nation's primary vote rising against both major parties.

Background

Australia is among several countries working out whether AI companies must license copyrighted material to train large language models. Anthropic has publicly warned that Australia's current copyright settings are blocking investment, and the major parties have been looking for common ground on AI regulation as public concern grows. The issue is also politically charged, with One Nation's strong polling pushing Labor and the Coalition closer together on policy fronts.

Key Perspectives

Labor and the Coalition: Want to attract major AI firms while compensating creative rights holders, but are wary of forcing individual licensing deals that the companies oppose and that could set a global precedent. Anthropic and global AI firms: Argue Australia's copyright laws are deterring investment in local operations. Greens and Senator David Pocock: Fear a bipartisan deal will short-change creators whose work is used to train AI models such as Claude.

What to Watch

  • Labor's formal position, expected once the government lands on a proposed solution, and whether the Coalition accepts it.
  • Whether Anthropic and OpenAI follow through on warnings about investment if the deal does not resolve copyright uncertainty.
  • One Nation's primary vote in coming polls, a key driver of the bipartisan urgency.
  • Any sign of legal action by international rights holders under Australian law.

Sources

Zotpaper

Written by software from the reporting listed above, scored by an automated standards desk, and published without a person reading it first. If something here is wrong, tell the editor and it will be put right.

How we workSubscribe