Australia at the AI crossroads: Why we must lead, not lag

In this month’s thought piece, the Director of the UNSW AI Institute, Dr Sue Keay, writes that Australia cannot afford to sit on the sidelines of artificial intelligence, as relying on foreign technology will undermine our sovereignty, security, and economic future. Dr Keay calls for urgent national investment in sovereign AI infrastructure, datasets, and public-private collaboration to ensure Australia shapes its own future.

Australia has never been short on ambition.

We built the Snowy Hydro Scheme, one of the world’s most audacious engineering projects, transforming our economy with renewable energy and irrigation.

We pioneered Wi-Fi through CSIRO, a technology now used in billions of devices worldwide.

We gave the world the cochlear implant and the Gardasil vaccine, changing lives across continents.

Dr Sue Keay, Director of the UNSW AI Institute. Credit: Supplied

This track record should remind us of what’s possible when Australia decides to lead.

Yet when it comes to artificial intelligence (AI), I keep hearing the same tired excuses: we’re “too small,” “too late,” or we should simply “buy it off the shelf.”

This kind of rhetoric isn’t just unhelpful, it’s dangerous.

AI is not another consumer gadget we can import.

It is the infrastructure of the future, shaping our economy, our national security, our democracy, and even our cultural values.

If we abdicate responsibility for developing our own capabilities, we are choosing dependency over sovereignty.

Sovereignty matters in a digital age

Australia’s track record of bold innovation—including Wi-Fi (pictured)—shows what’s possible when the nation chooses to lead. Credit: AdobeStock

Australians developing AI for Australians is not about vanity, it’s about agency. A sovereign AI capability ensures we can:

  • Protect our national security from foreign influence or manipulation.

  • Unlock our national datasets, built over decades, for innovation that reflects local conditions.

  • Retain and develop homegrown talent instead of losing it offshore.

  • Embed Australian values of fairness, ethics, and accountability into the technologies that will run our lives.

Optionality matters. If we don’t build now, the choice will be gone.

Our peers understand this. The United States, United Kingdom, Canada, South Korea, Singapore, and the EU are investing billions into sovereign AI infrastructure; public compute, curated datasets, and safety institutes, because they know markets alone won’t deliver what’s needed.

If we stand back, we don’t save money. We buy dependency, on foreign compute, foreign models, and foreign rules. And once locked in, those dependencies are almost impossible to unwind.

Learning from our own history

We’ve been here before. Government co-investment in aluminium and automotive industries in the mid-20th century wasn’t about “picking winners.” It was about building strategic capacity. Capacity that proved decisive in wartime.

By contrast, when we walked away from the car industry, we didn’t create efficiency, we hollowed out a complex capability, and with it slid to 93rd on the OECD’s economic complexity rankings.

Do we really want to repeat that mistake with AI, a technology that will shape every industry for decades to come?

The US, UK, Canada, South Korea, Singapore, and the EU are investing billions into sovereign AI infrastructure such as supercomputing facilities. Credit: AdobeStock

The Productivity Commission itself estimates AI adoption could add $116 billion to our economy every decade.

Studies suggest applying AI in key sectors; health, finance, education, mining, energy, could deliver a $200 billion annual boost to GDP and create 150,000 high-value jobs by 2030.

Fail to adopt, and we face a short-term opportunity cost of 1.4 percent of GDP, about $35.7 billion per year, by falling behind global early adopters.

The numbers are clear: abstaining isn’t strategy. It’s surrender.

What a practical plan looks like

Australia doesn’t need to reinvent the wheel, we need to act decisively on a focused national mission:

  • A national AI compute utility: shared, on-shore GPU and HPC facilities so researchers, SMEs, and agencies can compete with international peers.

  • Sovereign models and datasets: co-funded, safety-audited models trained on curated Australian data under strong privacy protections.

  • Procurement as a lever: government purchases in health, justice, and education to pull innovation into real-world use.

  • Clear guardrails: open standards, mandatory co-investment, stage-gates, and independent evaluation to keep innovation accountable and safe.

This is the formula other nations are already pursuing. It is how public investment catalyses private markets, spurring adoption, exports, and scale.

Given the platform, Australian business will respond with co-investment, rapid uptake, and global ambition. Our track record proves it: Cochlear and ResMed are global leaders born from public-private collaboration. AI can follow the same path, if we build the foundations.

Backing the next generation

AI is not a “nice to have.” It is the foundation of future industry and security. It is not a substitutable product; it is the bedrock upon which Australia’s productivity and prosperity can be built.

AI sovereignty is about backing the next generation. Credit: AdobeStock

As Dave Lemphers of Maincode (developers of an Australian sovereign LLM) reminds us: “AI sovereignty is not about technology, it’s about backing the next generation”.

The decision before us is stark. Either Australia meets this moment with vision and investment, or we consign ourselves to being consumers of someone else’s future.

This is not about nostalgia for lost industries. It is about leadership. Leadership to ensure our children inherit not just a thriving economy but the tools to shape it.

Australia has the talent, the data, the energy, and the democratic values to be a genuine AI nation. What we need now is the will.

Let’s not sit on the sidelines. Let’s meet the moment.

Learn more about the AI Institute.

Diane Nazaroff