When technology makes work better
In 2016, Geoffrey Hinton, the 'godfather of AI', made a stark prediction. 'We should stop training radiologists now', he said, suggesting that artificial intelligence would soon take over the job of reading scans.
It sounded plausible. Radiology looked like the perfect test case for automation: digital images, repeatable tasks, clear benchmarks. Algorithms such as CheXNet began outperforming panels of expert radiologists in spotting pneumonia. Companies launched hundreds of products to prioritise urgent scans, draft reports and detect disease. If any profession was about to be wiped out by machines, surely radiology was it.
And yet, the opposite happened. As researcher Deena Moussa notes, in the United States, radiology residencies are oversubscribed and salaries have soared. In Australia, the Royal Australian and New Zealand College of Radiologists currently lists more than fifty unfilled positions, while Glassdoor reports that the average radiologist earns multiples of the national average wage. Far from being displaced, radiologists are more productive and more valued than ever.
Why? Because AI tools do not remove the need for radiologists, they change the way radiologists work. Machines speed up routine tasks, so radiologists can focus on complex diagnoses and on conversations with clinicians and patients. Faster, cheaper scans mean doctors order more of them, which in turn creates more work for radiologists. Far from hollowing out the profession, AI has deepened its importance.
That is the optimistic story of innovation and productivity. Technology has the potential to make jobs better, not worse. It allows us to do more, not less. Productivity growth is not about working longer hours. It is about using ideas and innovation to make every hour more valuable. And it is not just about radiology. The same dynamic is at play across the economy: in energy, in agriculture, in advanced manufacturing.
That is the story I want to follow today. I will begin with the state of Australia's productivity challenge. I will then outline the government's agenda, framed around individuals, infrastructure and institutions. From there, I will turn to the role of research‑intensive universities - the places where we both generate these breakthroughs and train people to use them. Finally, I will offer some reflections on what more is needed if research and innovation are truly to turbocharge our nation's productivity.
Australia's productivity puzzle
For much of the twentieth century, productivity growth was the quiet engine of Australian prosperity. It allowed wages to rise, hours to fall and living standards to improve. But over the past 2 decades, that engine has sputtered. Both labour productivity and multifactor productivity have stalled.
The causes are many. There has been insufficient investment in productive assets and innovation. Business spending on R&D has lagged. It is harder to eke out productivity improvements in some of our fastest‑growing sectors - such as aged care and education, and tricky to properly measure productivity improvements in sectors such as healthcare. And the gap between frontier firms and laggards has widened, as too few companies adopt best practice.
This matters because when productivity growth is weak, wages stagnate and opportunities shrink. Choices about budgets become harder and intergenerational mobility slows. For Australia, lifting productivity is not just an economic task, it is a social imperative.
The progressive productivity agenda
If we are to reignite productivity growth, it must be through a broad‑based agenda that invests in people, builds the foundations for new industries, and ensures our institutions are fit for purpose. I think of this progressive productivity agenda as the 3 'I's: individuals, infrastructure and institutions.
The first is individuals. People‑powered productivity depends on whether Australians have the skills, adaptability and confidence to seize new opportunities. That begins with early childhood education, where quality is critical. It continues through schools, where data fluency, critical thinking and creativity are becoming almost as important as literacy and numeracy. And it extends into vocational education, universities and lifelong learning.
Education should encourage productive technological adoption. Consider the example of health: AI is reshaping radiology, but only radiologists who know how to work alongside machines can make the most of it. The same is true for accountants working with machine learning or farmers using satellite data.
The second arm of our productivity agenda is infrastructure. In the 20th century, it was railways, ports and electricity grids that lifted output. Today, it is broadband networks, renewable energy transmission and data centres. These investments are not only about connectivity or climate, they are the arteries of productivity - enabling ideas, power and people to flow more freely.
Digital infrastructure is a good example. Over the past decade, many communities shifted from dial‑up or patchy mobile connectivity to widespread access through the NBN network. Today, the vast majority of premises are in areas served by fixed or fixed wireless NBN - which underpins growth in telehealth, remote work, digital exports and more.
Energy infrastructure tells a similar story. The transition to clean energy is not simply an environmental choice, it is an economic one. Renewable power is now the cheapest form of electricity in history. By rewiring the nation and building a modern transmission system, we can attract energy‑intensive industries, from green hydrogen to advanced manufacturing. Productivity is about lowering costs and raising possibilities, and infrastructure is what makes that possible.
The third component is institutions. Good rules and incentives help shape whether innovation spreads or stalls. That is why the government is modernising merger laws to ensure markets remain dynamic and contestable, and why we are reducing non‑compete clauses that too often trap workers in jobs rather than letting them move where their skills are most valuable.
Institutions also matter in health research, where reforms to clinical trials will create a National One Stop Shop, standardised procedures and stronger governance. These changes will help promising treatments reach patients sooner, while cementing Australia's place as a destination for cutting‑edge research. Strong institutions like these provide the trust and certainty that allow good ideas to diffuse quickly through the economy.
Taken together, these 3 I's form a progressive productivity agenda. They are the scaffolding within which research and innovation can flourish. They ensure that when a new idea emerges, we have the people to use it, the infrastructure to spread it and the institutions to support its diffusion. And that is where universities come in, because they contribute directly to each of these dimensions.
The role of research‑intensive universities
Australia's research‑intensive universities are essential to productivity because you touch all 3 dimensions at once.
On individuals, you produce the graduates who will carry new skills into the workforce. You train the scientists who push the frontier, and the PhD researchers who will lead not only in academia, but in business, government and the community sector. The private benefits of a degree are substantial. Recent research from Tom Karmel at the Mackenzie Research Institute estimates large earnings premia for bachelor degrees: 53 per cent in science, 61 per cent in commerce, 65 per cent in arts, 68 per cent in engineering. Education also has a positive externality, benefiting co‑workers through peer effects and society through higher tax revenues.
On infrastructure, universities sustain the laboratories, libraries and data systems that underpin innovation. These are assets no single firm could build alone, but they make discoveries possible. Radiology's leap forward came from advances in machine learning and large annotated datasets. Similar breakthroughs in energy, agriculture or defence depend on the infrastructure that you maintain.
And on institutions, you embody the norms of open inquiry and collaboration that allow ideas to diffuse. You connect Australia to global research networks, bringing knowledge from abroad and sharing our own discoveries with the world. You act as hubs of civic and community life. You generate spillovers that reach far beyond the lecture theatre or the lab. Productivity depends not only on invention, but on diffusion, and you are among the strongest diffusion engines we have.
Within that landscape, the Group of Eight play a distinctive role. You educate around 1 in 3 Australian university students and account for about 70 per cent of Australian university research. Your graduates are disproportionately represented in leadership roles across business, government and science. Your laboratories, libraries and data assets form much of the backbone of national research capacity. And your international partnerships bring global knowledge to Australia while projecting our own expertise abroad.
But you cannot, and should not, carry the task alone. The contribution of research‑intensive universities is maximised when you are part of an ecosystem that includes innovative businesses, applied research institutes and public policies that encourage risk‑taking and diffusion. That is why business-university collaboration matters so much.
Right now, the higher education sector carries much of the nation's research load, while business R&D lags. That imbalance leaves us with a lopsided ecosystem. Innovation flourishes when universities and firms work together, when discoveries flow into prototypes, and when market challenges inspire new research.
Too many Australian ideas still fall into the 'valley of death' between discovery and commercialisation. The government is helping bridge that gap through the National Reconstruction Fund, Trailblazer programs and translational research initiatives. But collaboration also requires cultural change: more mobility of people between sectors, more joint appointments, and more investors willing to back early‑stage technology. Procurement can play a role too, by acting as an early customer for homegrown innovation.
Internationally, Germany's Fraunhofer Institutes show how applied research organisations can bridge the lab and the factory floor. We need not copy their model wholesale, but we should be willing to learn from their success in moving ideas into action.
Beyond collaboration, productivity also depends on cultural and institutional shifts. We need a greater tolerance for risk‑taking and for learning from failure. We need digital infrastructure that plays the role of electricity grids in earlier eras. We need a stronger culture of evaluation, so that resources are directed to what works. And we need to value PhDs not just as academics‑in‑waiting, but as assets to business, government and society.
At your best, you contribute to all of this. You foster the culture of experimentation. You sustain the infrastructure. You embody the institutional strength and global openness that allow ideas to spread. And you are where Australians learn - as radiologists have - that technology can make jobs more rewarding, not less.
The road ahead
The story of radiology is not about a profession replaced by artificial intelligence. It is about a profession transformed, made more productive, more valuable, and more rewarding. As economist Joshua Gans notes, if radiologists are the canary in the coal mine, 'the canary is well and truly alive'.
That captures the essence of productivity. It is not measured by the number of hours we put in, but by how much value we create in those hours. It is not about fewer jobs, but more interesting jobs. Productivity turns disruption into opportunity, replacing drudgery with discovery.
Australia's challenge is to apply that lesson across the economy. Universities, businesses and government must work together to ensure that research does not remain trapped in labs and journals, but spreads through every workplace and community. We need the people with the skills, the infrastructure to connect them, and the institutions to carry new ideas further and faster.
If we get this right, the payoff is more than numbers on a chart. It is Australians enjoying higher wages and shorter hospital waiting lists. It is faster commutes and cleaner energy. It is new industries that we have not yet imagined, and new opportunities for the next generation.
So let us see productivity for what it really is: not an economist's abstraction, but the secret sauce of national progress. It is how we make work more rewarding, communities more dynamic and the future more abundant.
