Top artificial intelligence executives told a parliamentary inquiry in Australia on Tuesday that their firms never sought to dictate local copyright rules. Meanwhile, creative industries warned lawmakers that weakening protections for AI model training would leave artists as unprotected victims.
Anthropic and OpenAI used appearances before the joint parliamentary hearing on artificial intelligence to defend their approaches to local creative works. Federal lawmakers are weighing competing demands from major technology developers, creators, and the country’s expanding data-centre industry following discussions on Tuesday morning centered on copyright licensing for AI training.
Anthropic Denies Policy Pressure While Warning Training Is Impossible Under Current Rules
Anthropic’s special envoy, Jeffrey Bleich, pushed back against suggestions that the company tried to force Canberra’s hand on intellectual property. We’ve just shared information about how the training works, and then discussed what its implications would be with respect to our capacity to operate here,
Bleich told the committee.
Bleich stated that Anthropic had never tried to dictate to Australia regarding copyright rules. Existing legal frameworks require AI developers to secure individual licenses for every piece of content across the internet, a requirement Bleich described as technically impossible.
Anthropic confirmed it will not train Claude models domestically until a viable payment regime exists, having previously warned that Australia risks missing out on a major portion of planned data centre investments unless regulations evolve. In September submissions to the parliamentary inquiry, Anthropic acknowledged that broad copyright exemptions lacked sufficient support from the government.
Anthropic accepts that broad exceptions from copyright …
Anthropic, via Firstpost
The company proposed a narrow conditional approval mechanism as an alternative. As an alternative, we believe that the government could permit AI model training through a narrow form of conditional approval,
its submission stated. The firm also noted openness to investment or other conditions designed to support the future of Australian creators and cultural endeavours if regulators require them.
OpenAI Apologizes for Government Website Breaches
I want to begin with an apology. During internal training and evaluation, our models accessed Australian government websites in ways they were not directed to. That should not have happened.
Jason Kwon, Chief Strategy Officer at OpenAI, via The Guardian
Both tech firms tied their regulatory requests to active local infrastructure projects, including OpenAI’s offtake agreement with NextDC for a planned facility in Sydney and Anthropic’s partnership for a Queensland data centre. OpenAI urged lawmakers to adopt a balanced copyright framework that allows models to learn from publicly available information while establishing collaborative opportunities for rightsholders.
Creative Sector Representatives Advocate for Equitable Remuneration Scheme
Creative sector representatives pushed back hard against the technology companies during the hearings.
Acting chief executive Adam Portelli of the Media Entertainment and Arts Alliance advocated for an equitable remuneration scheme, proposing fees across the technology stack to fund direct payments for media workers alongside a cultural fund.
The Joint Select Committee on Artificial Intelligence continues hearing testimony through Friday and is scheduled to deliver its formal report in November.