On 15 July 2026, the Albanese government announced $3.6 billion for artificial intelligence. There were data centres, enterprise grants, safety frameworks, sovereign capability commitments, and all the language that signals a government that has paid attention to the briefings. It’s a serious package. It’s a response to a real moment. And when I read it, I couldn’t shake a question that the press release doesn’t answer. Who is this for? That’s not a political question. I’m not interested in which side of the aisle gets credit or whether the numbers are right. I’m asking something more fundamental: is this a strategy for AI in Australia, or a strategy for Australian AI? They’re not the same thing.

A Strategy for Australian AI Is Not a Strategy for AI in Australia

A strategy for Australian AI is about infrastructure, enterprise, sovereign capability, and governance. It’s about making sure Australia can compete at the technology layer: data centres, compute, R&D spend, business adoption. That’s what this announcement is. It’s coherent. It’s defensible. But it’s essentially a plan for industry. A strategy for AI in Australia would ask something different. It would ask what AI means for the Australian who works in a warehouse, teaches in a regional school, runs a small business in a town that doesn’t have a tech hub, or makes art that nobody outside of Melbourne has ever heard of. It would ask what access to AI capability means for someone who can’t afford an enterprise subscription. It would also ask what happens to Australian culture: our specific, distinct, hard-won cultural voice, when AI reshapes what gets amplified globally and the institutions that carry that voice have no resources to engage. That question is nowhere in this announcement. Here’s what I keep coming back to when I think about moments like this: governments regulate what they fear and fund what they want to own. This announcement does both. It puts guardrails around AI risk (safety frameworks, ethical standards) and it invests in AI infrastructure (data centres, enterprise adoption). What it doesn’t do is treat AI capability as something Australians should all have access to, the way we once, imperfectly and with great argument, tried to treat broadband.

The Access Question Nobody’s Asking

The NBN debate was fractious. The rollout was messy. But the underlying principle was worth arguing for: this technology is too important to leave only to those who can already afford it. I haven’t heard that argument being made about AI. I don’t see it in this strategy. And that matters, because the gap between people who can use AI and people who can’t is already widening. Not just in income terms. In literacy terms. In confidence terms. In terms of who feels like the future is being built with them rather than around them. A strategy that builds infrastructure at the top and sets rules in the middle isn’t going to close that gap. It’s going to formalise it.

What Other Countries Are Actually Doing

Last month, Canada launched an AI strategy called, simply, “AI for All.” Not a tagline. The actual name. Its centrepiece is a National AI Literacy Initiative: free AI training for every Canadian, classroom kits for thousands of educators, AI tools for every post-secondary student from arts to medicine, and specific programs to bring small businesses along. Finland has been running free national AI literacy courses for years, deliberately aimed at the entire population, including the elderly. Singapore’s strategy is built on an explicitly different premise: whole-of-economy, whole-of-society, embedding AI across every sector from healthcare to education to logistics, not just the technology industry. All three made the same decision: that AI capability is a public resource, not just an industry investment. That layer is absent from Australia’s announcement entirely. There’s a second thing this strategy doesn’t address, and I think it’s worth naming directly: brain drain.

The Brain Drain Problem

Australia has been exporting its best AI researchers for years. They’re at DeepMind, at Google, at Anthropic, at OpenAI, doing their best work on someone else’s payroll, in someone else’s time zone, building someone else’s future. Does this announcement give a single compelling reason for any of them to come home? Or for a student finishing a machine learning PhD in Melbourne to stay, rather than accepting the offer that’s already in their inbox from a US company with ten times the compute access and twice the salary? I don’t see it. The announcement mentions sovereign capability. But sovereign capability isn’t built by funding data centres. It’s built by creating conditions where the people who are brilliant at this work want to do it here, with Australian institutions, on problems that matter to this country. That requires a different kind of investment: in research careers, in university partnerships, in the kind of cultural infrastructure that makes a place feel like somewhere worth building a life.

The Question That Wasn’t in the Briefing

I’m asking whether, before the press releases went out and the funding commitments were made, anyone in the room asked the question that didn’t appear in the briefing document. Not “how do we build Australia’s AI capacity?” but “what kind of country do we want AI to make possible?” Those are different questions. The first one gets you a funding package. The second one gets you a direction. This announcement is a response to what’s already happening. AI is here, the world is moving, we need to catch up. I understand that logic. But foresight isn’t about catching up. Foresight is about choosing where you want to arrive before you decide what road to take. Australia is choosing a road right now. And the road we’re on leads to capable infrastructure, enterprise efficiency, and governance frameworks that the top end of town can navigate. That’s not nothing. But I’d want to know what road leads to a country where an artist in Hobart, a teacher in Broken Hill, and a first-generation Australian in Dandenong all feel like AI is working for them, not just around them. An inhabitable AI future for Australia isn’t measured by data centre capacity. It’s measured by whether the small business owner in Geelong has access to the same tools as the enterprise in Sydney. Whether the kid starting school this year, who will graduate into a world where AI is as fundamental as reading, is being equipped for that world rather than just introduced to it. Whether the researcher who could be at DeepMind is choosing instead to work on something that matters here. Whether Australian stories and Australian ways of seeing the world are still present in a global AI diet that will otherwise be shaped almost entirely somewhere else. That piece of the strategy is still blank. And that’s the piece worth demanding.

Frequently Asked Questions

What did Australia announce for AI in 2026?

The Albanese government announced $3.6 billion for artificial intelligence in July 2026, covering data centres, enterprise grants, safety frameworks, and sovereign capability commitments. It also established an Office of AI within the Department of the Prime Minister and Cabinet to coordinate national AI policy.

What is the difference between a strategy for Australian AI and a strategy for AI in Australia?

A strategy for Australian AI focuses on infrastructure, enterprise, sovereign capability and governance — making Australia competitive at the technology layer. A strategy for AI in Australia asks what AI means for every Australian: the warehouse worker, the regional teacher, the small business owner, the artist. The 2026 announcement is primarily the former.

How does Australia compare to other countries on AI literacy?

Canada launched an AI for All strategy with a National AI Literacy Initiative providing free AI training for every Canadian. Finland has run free national AI literacy courses for years across the entire population. Singapore’s strategy embeds AI across every sector. Australia’s 2026 announcement does not include a comparable public AI literacy or access component.

What is missing from Australia’s AI strategy?

The main gaps are: a public AI literacy and access program comparable to Canada’s AI for All or Finland’s national courses; a plan to retain AI researchers rather than losing them overseas; and a clear vision for what kind of country Australia wants AI to make possible — not just infrastructure investment targets.

What should Australian business leaders do in response to the AI announcement?

Three things: map your AI exposure including where AI decisions are being made without governance; develop an explicit position on how work is allocated between humans and AI; and track the energy question, since data centre power demands will affect any organisation that competes for grid capacity.

Choose Forward.