OpenAI and Anthropic urge Australia to loosen AI training copyright ban

OpenAI and Anthropic responded to a September parliamentary inquiry by requesting that Australia reconsider its blanket ban on using domestic creative content for AI training. While accepting that broad copyright exceptions have been rejected, they propose a narrow, conditional approval system that could include requirements such as investment in local creators or cultural initiatives. Anthropic expressed willingness to work under such conditions, and OpenAI advocated for a balanced copyright regime that still offers rights-holders collaboration opportunities.

The companies linked their proposals to major infrastructure commitments, noting OpenAI's offtake deal with data-centre developer NextDC in Sydney and Anthropic's partnership on a Queensland data centre. The Joint Select Committee on Artificial Intelligence is slated to release its report in November, putting pressure on policymakers to balance AI development with creators' interests.