The US government can shut off access to artificial intelligence at will. What does this mean for Australia? Last Friday, the US-based AI company Anthropic received an export control directive from its government. The company was ordered to block access to two of its most capable models, Fable and Mythos, for all foreign nationals. Within hours, Anthropic cut off access for users worldwide, including researchers, clinicians, and analysts in Australia. This happened without warning or a backup plan.
Why Did This Happen?
The directive's foreign national criterion is based on citizenship. However, Anthropic and other cloud-based AI providers only know users' locations, not their citizenship. Consumer AI services lack effective mechanisms to verify citizenship. Even location filters can be bypassed using tools like VPNs. There is no way to enforce controls based on user nationality. A system that blocks a foreign national in the US but allows a US national in Australia is not currently available. The US government directive was a geopolitical signal—an indication rather than an enforceable control. The only way Anthropic could comply was to shut down access for everyone.
A Question of Sovereignty
Australian universities, government agencies, health systems, and industry have deeply integrated US-hosted frontier AI into their operations. Advanced analytics platforms, AI-assisted research tools, and productivity infrastructure built on models like Claude or GPT-5 operate with an implicit assumption that access will persist. The same is true for ubiquitous systems like Microsoft Teams or Google Workspace, but as we see, this assumption may not hold for AI. The Anthropic shutdown reminded us that access and control sit entirely within US jurisdiction, regardless of where Australian users are. It represents a failure of data and technology sovereignty: our ability to operate without permission from other nations.
Australia’s Exposure
The Anthropic shutdown is the first legally enforceable export control targeting a software-level AI system. It will not be the last. Any frontier AI model hosted in the US—by Anthropic, OpenAI, Google DeepMind, or Meta—is subject to US export control law. The precedent set this week extends to all of them. Some debate exists about whether export controls apply only to physical exports and not remote access to models housed in the US. Some experts suggest the order may be challenged on those grounds. If so, the US government would likely change the regulations. The US has previously used technology export controls as a geopolitical instrument, especially with chips and semiconductors. Even allies like Australia are not exempt. The Anthropic order should not have been a surprise. Voices in Australia’s research community, including Jon Whittle, former head of CSIRO Data61, had been publicly warning about this scenario for over a year. To prepare for the future, Australia needs a strategy for its own sovereign AI. This cannot be a distant aspiration: it must be an operational plan with named owners, timelines, and budget.
How Does Australia Achieve AI Sovereignty?
A true sovereign AI strategy requires four elements. First, data sovereignty: data physically stored in Australia and subject to Australian law. Second, compute sovereignty: in-country data centers under Australian control. Third, AI model sovereignty: AI capability not dependent on a foreign provider. Fourth, policy sovereignty: the ability to set its own rules, rather than inheriting another country’s export controls or safety regulations by default. Australia already has examples of sovereign data and compute, such as Macquarie Government, Vault Cloud, and AUCyber. Building this capacity will require huge, sovereign data centers on Australian soil. However, that build-out will only gain public support if it comes with binding commitments on renewable energy sourcing, water-efficient cooling, and transparent community consultation in regions where data centers are proposed, such as Bundey in South Australia. For Australian residents, this sovereignty means not depending on AI tools and systems that another nation can switch off, restrict, or alter at will.
Developing sovereign AI models will be a challenge. However, it does not necessarily mean building new systems from scratch. Switzerland’s Apertus model, released in September 2025, is an excellent example of what open-source approaches can produce. It also shows how large language models can comply with regulations like the European Union’s Artificial Intelligence Act. Beyond this, sovereign capability will require training future experts. We can expect increased demand for people who can build the technology, deploy and manage it, govern it, and integrate it into existing systems.
The Regulatory Dimension
The EU is developing AI policy frameworks from which Australia can learn. It is establishing reciprocal access rights with technology partners, doing institutional contingency planning, and putting money into the process. Several European countries are moving rapidly. France has just announced an additional €655 million (A$1 billion) for national AI capabilities, explicitly referencing last week’s events. This builds on roughly €2.5 billion of investment since 2018. Sovereign AI does not mean reinventing the frontier model from scratch. It can mean public, transparent, nationally governed infrastructure built on open foundations. At a minimum, AI sovereignty requires ensuring we are not too dependent on single AI suppliers or even suppliers based in a single jurisdiction. The goal should be that no single government or jurisdiction can unilaterally cut off Australian residents and institutions.
