Anthropic Presses Australia for Copyright Reform in AI Negotiations

Anthropic chief executive Dario Amodei lobbied Australian officials for changes to copyright law as the company explored a major AI investment in the country, according to Reuters via Asharq Al Awsat.

Amodei met with Australia’s Treasurer Jim Chalmers in April to discuss plans to expand into the Australian market, which included building data centers. Briefing notes released under freedom of information law showed that Amodei raised concerns about barriers to AI training, specifically those linked to copyright licensing.

The documents stated that Anthropic’s potential investment in developing AI model infrastructure would depend on regulatory clarity around copyright. In the United States, Anthropic has maintained that AI training qualifies as fair use, which does not require the consent of rights holders. Australian officials, however, described the issue as unresolved and indicated the country would not introduce a text and data mining exception to its copyright law.

Under current Australian law, AI companies must obtain licenses from copyright holders to use their material. Officials also noted that Anthropic described difficulties in identifying smaller rights holders when seeking licenses. The government continues to consult with industry and creative sector stakeholders on the matter.