Copyright Law Becomes Key Battleground for AI Investment in Australia

Summary

Australia's ambitions to become a global AI powerhouse are now being tested by a fundamental question: who owns the rights to data used to train AI models? According to internal government documents revealed by the ABC, copyright law has become a major sticking point for companies like Anthropic, whose Claude AI is one of the world's most advanced AI systems. Anthropic indicated that major investment in Australia—potentially worth tens of billions of dollars for data center buildouts—is contingent on achieving "clarity of copyright settings."

The conflict stems from a 1968 copyright law that predates the digital age, let alone generative AI. Under current Australian law, the act of training an AI model on copyrighted material could involve multiple stages of infringement: collecting material, organizing copies, testing the model, and deploying it. Professor Kathy Bowrey of UNSW notes that Australian copyright law offers developers fewer defenses than US law, where courts can evaluate "fair use" claims. Meanwhile, content creators—writers, musicians, artists, and publishers—argue they deserve control over their work and fair compensation for material absorbed by AI companies to train their systems.

Prime Minister Anthony Albanese is set to announce the government's AI strategy this week, with copyright clarity reportedly central to the investment equation. The stakes are enormous: Australia's bid to secure world-leading AI infrastructure, jobs, and economic growth hinges on resolving a legal framework that predates even the moon landing.

• Prime Minister Albanese's forthcoming AI strategy speech is expected to address copyright clarity as critical to Australia's AI competitiveness