Anthropic and OpenAI (discover more) Ask Australia to Ease AI Training Restrictions
Anthropic and OpenAI have approached Australian lawmakers, requesting them to reconsider the country's ban on AI training with local content. According to Reuters, the companies have submitted their proposals to a parliamentary inquiry focusing on AI laws.
Background
In October 2025, the Australian government announced it would not include a text and data mining exception in its copyright legislation. This decision effectively prevents AI developers, primarily based in the US, from training models on Australian creative works without prior permission from copyright holders.
Companies' Proposals
- Anthropic: Acknowledges the ruling against broad copyright exceptions and suggests a 'narrow form of conditional approval' for model training. They are open to conditions that support Australian creators and cultural works.
- OpenAI: Requests a balanced framework that allows models to learn from publicly available information while providing rights holders with opportunities to collaborate with AI companies.
Both companies tie their proposals to planned data centers in Australia. OpenAI has a power purchase agreement for a Sydney site, and Anthropic is a partner in a Queensland data center project.
Connection to Copyright Discussion
Anthropic's CEO, Dario Amodei, reportedly discussed copyright with Treasurer Jim Chalmers in April, highlighting the company's investment dependency on copyright clarity.
The creative industries have also expressed their stance, with Green senator Sarah Hanson-Young stating that AI companies should "pay for [copyrighted material] just as everyone else does."
Anthropic has faced similar issues in the US, settling a class action for $1.5 billion over pirated books used in its training data.
The submissions were made to a joint select committee on AI, established by both houses of the Australian parliament in August, as the country prepares to introduce more comprehensive AI and data center regulations next year.