OpenAI and Anthropic have submitted proposals to the Australian parliamentary committee regarding the use of copyrighted content for AI training. Both companies oppose a regime where each piece of material must be licensed separately. Anthropic suggests a narrow conditional permission model, with commitments to local authors and cultural projects. OpenAI advocates for a balanced model: training on publicly available data with the possibility of separate agreements with rights holders.
The companies are considering copyright issues in the context of developing AI infrastructure in Australia. OpenAI, in collaboration with NEXTDC, is preparing a project for the S7 computing complex in Sydney, having signed a memorandum for joint development. Anthropic is leasing capacity from the Zerra DC data center in Queensland, which has a stated capacity of up to 2.16 GW, with commissioning planned for 2027; the campus is located about 250 km from Brisbane. Representatives from creative industries oppose expanding exceptions.
The parliamentary committee, established on August 20, is examining the impact of AI on the economy, competitiveness, and intellectual property legislation. The final report is expected to be ready by November 30. On September 22, the head of the Reserve Bank of Australia, Michelle Bullock, stated that she does not see compelling evidence of productivity growth due to AI and suggested that the investment boom surrounding the technology could turn out to be a bubble. In her opinion, the construction of data centers is currently increasing demand and could sustain inflationary pressure.
Related articles
