The Australian Strategic Policy Institute has moved to plant a flag in one of the most contested corners of the technology debate, launching a dedicated Artificial Intelligence and Security Program aimed at shaping how Australia thinks about the national security implications of a technology that is advancing faster than most policy can keep up with.
The Canberra-based think tank, which has spent more than two decades weighing in on defence, cyber and strategic policy, said the new stream would concentrate on where artificial intelligence intersects with security, from military applications and intelligence work through to the resilience of the systems and supply chains that increasingly depend on it. You can read the institute’s own announcement here.
Why a think tank is stepping in now
The timing is not accidental. Governments around the world are scrambling to work out how to regulate, adopt and defend against artificial intelligence at the same time, and Australia is no exception. Canberra has been drafting its own guardrails for high-risk uses of the technology, the Australian Defence Force has been openly canvassing autonomous systems and AI-assisted decision-making, and the intelligence community has flagged both the opportunities and the threats that come with machine learning at scale.
Into that gap steps ASPI, betting that a permanent research capability focused squarely on AI and security will give it a durable voice in a conversation that has so far been dominated by a mix of commercial vendors, academic researchers and overseas policy shops. The institute has built its reputation on producing accessible analysis that lands in ministerial offices and newspaper columns alike, and the new program is clearly intended to do the same for a subject that many decision-makers find genuinely difficult to grasp.
The remit, as ASPI has framed it, is broad. It takes in the defence and offence dimensions of the technology, the way authoritarian states are using AI for surveillance and influence operations, the security of the data centres and semiconductors that underpin the whole edifice, and the harder questions about how much of this capability a middle power like Australia can realistically build and control on its own.
Two ways to read the move
Supporters of the initiative argue that it fills a real and growing hole. Australia has plenty of technical talent working on artificial intelligence, but comparatively little institutional muscle devoted to the strategic and security consequences of the technology. A dedicated program, on this view, gives policymakers a local source of analysis rather than leaving them reliant on research produced in Washington, London or Silicon Valley, where the assumptions and interests do not always line up with Australia’s. For a country trying to hold its own inside AUKUS and to build what officials increasingly call sovereign AI capability, having credible domestic expertise matters.
Critics, though, will read the launch through a more sceptical lens. ASPI has long attracted debate over its funding, which draws on the Department of Defence, foreign governments and, at various points, technology and defence companies. Detractors argue that this mix can shape the questions a think tank chooses to ask and the answers it tends to reach, and that an AI and security program risks amplifying a hawkish framing of the technology at exactly the moment when calmer, more independent analysis is needed. The institute has consistently rejected that characterisation and points to the transparency of its funding and the independence of its researchers, but the argument is a familiar one in Canberra and it will follow the new program around.
There is a third perspective worth holding onto as well, one that sits outside the think tank ecosystem entirely. Civil society groups and privacy advocates have warned repeatedly that framing artificial intelligence primarily as a security problem can crowd out the equally pressing questions about rights, bias and accountability. How ASPI’s program balances the hard-edged strategic material against those broader public interest concerns will say a lot about whether it becomes a genuinely useful reference point or simply another voice in the defence chorus.
What it means for Australia
For Australian readers, the significance is less about a single organisational announcement and more about what it signals. The country is quietly assembling the institutions of an AI security state, from the government’s work on high-risk AI rules through to the ADF’s growing appetite for autonomous systems and the recent surge of investment in local data centre and AI factory infrastructure. A think tank standing up a permanent program on the subject is a marker that this is now treated as a standing national security issue rather than a passing curiosity.
It also lands at a time when the economics of artificial intelligence are becoming a mainstream Australian concern. Billions of dollars are flowing into AI infrastructure on home soil, the technology is being blamed and credited for shifts in the labour market, and the major banks and miners are racing to deploy it. Security sits underneath all of that. The data centres need protecting, the models need to be trusted, and the supply chains that deliver the chips run through a handful of geopolitical chokepoints that Australia does not control. A program that connects those dots for policymakers could prove genuinely valuable, provided it earns the independence it will need to be believed.
What’s next
The real test will come with the output. Think tank launches are cheap; sustained, rigorous and genuinely useful research is not. Watch for the program’s first substantial publications, the calibre of the people it recruits, and whether it can speak credibly to both the defence establishment and the broader public that will live with the consequences of these decisions. Watch, too, for how transparent it is about who is paying for the work, because in this field trust is the whole currency.
If ASPI gets the balance right, Australia gains a sharper domestic lens on one of the defining strategic questions of the decade. If it does not, the program risks becoming one more institution talking about artificial intelligence rather than meaningfully shaping how the country handles it.
