Is Australia taking AI seriously enough? I was thinking about this over the weekend, prompted by an AI Action Plan released by America that reads like a moonshot.
It is big, fast and unapologetically strategic. Three moves stand out: accelerate innovation through open-source models, light-touch regulation and massive reskilling; build hard infrastructure through subsidised chip fabs, fast-tracked data centres, secure clouds and grid upgrades; and treat AI as foreign policy.
Where does that leave Australia? We have a mandatory Responsible AI in Government policy, an AI safety voluntary standard, and more than 1,500 AI businesses with fast recruitment.
But relative to the US plan we still trail on scale, compute and power, skills, AI diplomacy, and R&D leadership. We are still an AI-taker, not yet an AI-maker.
America is running. Australia is taking a leisurely jog. To achieve real sovereign AI, Australia should be matching the American pace on compute, energy and skills, not just ethics.
Five quick wins: supercharge local R&D and open-source strategies with matched public-private funding; fast-track critical infrastructure approvals and energy supply to data centres; establish a national AI workforce mission between universities, TAFEs and industry; use trade and security levers to embed Australian standards in Indo-Pacific supply chains; and build public trust through real productivity wins across health, defence and climate.
Bottom line: Australia can transition from AI-taker to AI-maker, but time is running out.