Key facts
- Greens Senator David Shoebridge believes Australia should develop its own AI capabilities.
- Shoebridge warns that relying on US-owned tech stacks leaves Australia open to coercion.
- He cited statements from billionaire AI owners indicating a political project favoring hard-right politics.
- Shoebridge advocates for a model of independent statutory AI ownership with community and business buy-in.
- He takes seriously warnings from AI insiders about potential existential risks.
- Shoebridge calls for Australia to model a world-best regulatory scheme for AI and seek international cooperation.
David Shoebridge, the newly appointed AI spokesman for Australia's Greens party, has outlined a vision for the nation to develop its own sovereign artificial intelligence capabilities rather than relying on foreign-owned technology. Appointed on September 7, Shoebridge expressed concerns that an over-reliance on US-owned tech stacks, from chips to AI models, poses significant risks to Australia's control over its culture, government, and economy.
Shoebridge pointed to countries like South Korea and France, which are pursuing national AI alternatives, as models for Australia. He suggested a mixed public-private ownership model, potentially an independent statutory body with community and business buy-in, as the foundation for a national compute and large language model infrastructure. He also highlighted statements from some US tech billionaires that suggest a political agenda aligned with hard-right politics and the US military, which he views as a cause for concern.
Addressing existential risk warnings, such as a senior Anthropic employee's estimate of a greater than 10% chance of AI causing human extinction within a decade, Shoebridge stated that such concerns from well-informed insiders must be taken seriously by politicians. He called for Australia to develop a leading regulatory framework for AI and to push for international collaboration on guardrails, even in the absence of US buy-in. He suggested that cooperation to create protective markets from US-driven AI could be a necessary dual project.
Shoebridge criticized the current Australian government's approach, which he described as a "patchwork mess of oversight" scattered across multiple ministers and agencies. He argued that the Prime Minister's establishment of a direct portal for industry access to his office, rather than focusing on a coherent regulatory framework, represents a "do-nothing model." He proposed the creation of a dedicated AI minister with a focused departmental structure to provide the necessary leadership for this monumental task.