Australia's central bank governor has cast doubt on the economic promise of artificial intelligence, telling a Sydney audience that the technology may yet turn out to be a speculative bubble and that there is still no evidence it is making the country's economy more efficient.
Reserve Bank of Australia governor Michele Bullock made the remarks at a CEDA event on Tuesday, days before an expected interest rate rise. Her caution stands in sharp contrast with the optimism emanating from Canberra, where the Albanese government has positioned AI as the answer to Australia's flatlining productivity.
A bubble risk central banks are watching
Bullock said a disorderly sell-off in technology valuations could spill over into broader economic activity, and confirmed that the possibility is on the radar of monetary authorities well beyond Australia.
"All central banks are a little bit worried about that," she said. "Some people think it's a bubble, some people don't. I don't have a particular view one way or the other, but it's a risk that I think we're watching."
The governor's remarks landed against a backdrop of jumpy markets. Resurgent oil prices and lingering inflation have been weighing on sentiment, yet American technology stocks rallied sharply on Monday after Meta unveiled its new Muse AI agent, sending the company's shares up 11%. That split — nervousness about stretched valuations alongside euphoria over each new model release — is precisely the dynamic Bullock described as a risk rather than a certainty.
Adoption without a productivity dividend
Central bankers, Bullock said, have broadly treated AI as the "great white hope" for lifting the productive capacity of their economies. The difficulty is that the hoped-for improvement has not shown up in the numbers yet. What has materialised instead is a wave of spending on AI tools and datacentre construction that is contributing to inflationary pressure rather than offsetting it.
To illustrate the gap between enthusiasm and measurable gains, Bullock pointed to research from the Reserve Bank's South Korean counterpart. That work found that employees who adopted AI tended to deliver the same volume of output while working about 1.5 hours less each week.
"While people fiddle around and try and figure out what to do with this new technology, productivity actually can decline," she said. "But once we reimagine our business processes… then you might see productivity take off."
The implication is that AI may need to be accompanied by deep organisational change — redesigned workflows, retrained staff and rethought business models — before it translates into the kind of economy-wide efficiency gains that governments are banking on. Until that happens, the investment surge itself functions as demand-side stimulus, which in an economy already battling price pressures can look more like a cost than a benefit.
The intergenerational report's AI bet
Bullock's scepticism is notable because it arrives on the same day the federal government released its intergenerational report, a document whose long-range projections rest heavily on AI delivering what it calls a "profound" boost to the economy.
The report forecasts that Australians' inflation-adjusted economic activity per person will climb from $99,200 today to $157,300 by 2066, an outcome that depends on productivity growth recovering. Treasurer Jim Chalmers, speaking on the Guardian's Australian Politics podcast, described AI as "the most transformative thing that will happen in our lifetime."
The projections assume productivity returns to its long-run trend of improving by 1.2% a year — a pace many economists regard as unlikely given recent performance. Under a weaker scenario in which productivity grows by just 0.8% annually, the report's own figures show per-person activity reaching only about $136,600, a materially smaller improvement.
The tension between the two visions is now a defining feature of Australia's economic debate. One branch of officialdom projects decades of AI-fuelled gains; the other, charged with setting interest rates month to month, says it has yet to detect them at all.
Housing, rates and the inflation squeeze
Bullock also addressed the property market, saying the recent slump in house prices had been deeper than others in Australia's recent economic history. The comment adds another layer of complication for policymakers, who must weigh softer household wealth against price pressures that have yet to be fully tamed.
With an interest rate rise anticipated next week, the governor's message was that the near-term inflation problem remains live. AI-driven capital expenditure is part of that story: datacentres consume enormous amounts of electricity, land and specialised labour, all of which can push costs higher in the short term even if the long-term payoff eventually arrives.
Why the timing matters
The episode captures a broader global moment in which the technology industry, investors and governments have collectively bet enormous sums on AI's capacity to remake the economy, while the statistical agencies and central banks that measure actual output have yet to record the transformation.
That does not mean the bet is wrong. Bullock's framing suggests a lag rather than a dead end: the productivity gains may come once firms stop experimenting and start restructuring. But it does mean the transition period can be uncomfortable — weaker measured productivity, higher prices and inflated asset values all at once.
For Australia, the stakes are unusually high. The intergenerational report effectively mortgages the country's future living standards to an AI-driven productivity revival. If that revival is slower or smaller than assumed, the fiscal arithmetic underpinning the projections weakens with it, and the burden shifts back to policy choices that are far harder to make than a forecast.
Bullock did not predict which way the question will resolve. She simply noted that the evidence is not yet in — and that central banks, whatever their hopes, will keep watching the data rather than the hype.
This article is based on reporting by The Guardian. Read the original article.
Originally published on theguardian.com