Anthropic and OpenAI have jointly approached Australian authorities, urging the government to relax restrictions that currently limit how AI companies can use locally produced content for training their models. The request, reported by The Next Web, puts both firms at the center of a growing international debate over who controls the data that shapes artificial intelligence systems and under what conditions it can be used.
What the Companies Are Asking For
Both companies argue that existing Australian rules create an uneven playing field compared to jurisdictions with more permissive copyright frameworks, such as the United States and parts of Asia. Their submissions to Australian regulators contend that restrictions on training data access could slow AI development and put Australia-based researchers and businesses at a competitive disadvantage. Anthropic has been particularly active in international policy engagements of this kind, having recently sent its leadership to engage with world governments at the highest levels alongside OpenAI and Google.
Key Facts
- Anthropic and OpenAI have formally submitted lobbying requests to Australian regulators.
- Australia currently restricts the use of locally produced content for AI model training without explicit consent.
- Both companies argue these rules limit their ability to compete globally.
- The move has drawn criticism from Australian publishers, journalists, and rights holders.
- Australia joins a growing list of countries reviewing AI training data policies.
The request comes at a moment when both companies are expanding aggressively beyond the United States. Anthropic recently hit a $965 billion valuation, and the pressure to sustain that growth means finding new data sources and new markets. Australian content, ranging from news archives to government records and academic work, represents a potentially significant training resource that the companies currently cannot access freely.
Australia risks becoming a secondary market for AI development if its regulatory settings are out of step with global norms.Anthropic, submission to Australian regulators (paraphrased)
Critics Push Back Hard
The proposal has not been well received by Australian content creators and media organizations. Publishers and journalists argue that allowing foreign AI firms to train on their work without compensation or consent would amount to a form of state-sanctioned intellectual property extraction. The concern mirrors debates playing out in the European Union, the United Kingdom, and the United States, where courts and legislators are still working through the legal boundaries of AI training data. Those following the G7 AI regulation discussions will recognize the pattern: AI firms lobbying hard ahead of regulatory frameworks being locked in place.
Australian regulators now face a familiar dilemma. Tightening rules risks alienating major AI investors at a time when governments worldwide are competing to attract the sector. Loosening them risks setting a precedent that weakens creators' rights and hands significant leverage to a handful of large American companies. Neither option is without political cost, and the outcome could influence how other mid-sized economies approach similar questions.
The Broader Regulatory Picture
This lobbying effort is part of a wider pattern of AI companies engaging proactively with governments before legislation solidifies. The strategy reflects a lesson learned from the social media era: that it is far easier to shape rules before they are written than to challenge them afterward. For those tracking the latest Claude AI news, it is worth noting that Anthropic's policy engagement has grown considerably alongside its commercial footprint, with the company now treating regulatory affairs as a core business function rather than an afterthought.
Australia has not yet signaled which direction it will move. Consultations are ongoing, and the government has shown willingness to take a cautious approach to AI governance generally. Whatever decision emerges is likely to be closely watched by both rights holders and AI developers in other Commonwealth countries considering similar frameworks.
