OpenAI and Anthropic have urged Australia to reconsider its current approach to the use of copyrighted creative material for training artificial-intelligence models, Reuters reported on September 22.
Australia has said it plans to introduce a broad set of AI and data-centre rules from next year and has ruled out a broad copyright exemption that would allow AI companies to train models on protected Australian content without existing legal constraints.
In submissions to a parliamentary inquiry, Anthropic said a narrow form of conditional approval could be considered if it included safeguards or investment requirements supporting Australian creators and cultural activity. OpenAI called for a balanced copyright framework that would allow models to learn from publicly available information while giving rights holders opportunities to collaborate.
These are policy positions advanced by the two companies. They do not amount to an Australian government decision to change copyright law.
The debate also involves authors, publishers and other creators, many of whom oppose allowing commercial AI training on their work without consent or compensation.
Both companies linked the issue to planned Australian infrastructure investment. OpenAI has an offtake agreement connected to a planned NextDC data centre in Sydney, while Anthropic has been named as a partner for a data-centre project in Queensland.
Australia's Joint Select Committee on Artificial Intelligence is due to report in November, meaning the policy debate remains open and no final change has been announced.