RBA Governor Warns AI May Be a Bubble With No Productivity Payoff Yet

Reserve Bank of Australia Governor Michele Bullock said on September 22, 2026, that artificial intelligence could be a market bubble and that there is no evidence yet it is making the Australian economy more efficient.

Speaking at a Committee for Economic Development event in Sydney, Bullock said a disorderly slump in technology valuations could damage economic activity. The remarks were published by the RBA as a Fireside Chat at the Committee for Economic Development featuring Michele Bullock RBA.

"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," Bullock said The Guardian. She said all central banks are "a little bit worried" about AI valuations.

Bullock described AI as "the great white hope" for productivity. On September 22, 2026, she said: "AI is the great white hope to improve productivity." She added that the RBA and overseas counterparts had seen little sign of that effect so far. Adoption costs and datacentre investment were adding to inflation instead of reducing it.

She cited research from South Korea's central bank finding employees adopting AI produced the same output but worked 1.5 hours less each week. That distinction matters. Output per hour rose. Aggregate efficiency did not.

The comments connect directly to Australia's intergenerational report. That report projected inflation-adjusted output per person would rise from $99,200 in 2026 to $157,300 by 2066, relying on a "profound" boost from AI. It assumed productivity growth of 1.2% a year. At 0.8% growth, per-person activity would reach only $136,600 by 2066.

Bullock also warned AI could lead to a difficult period with higher unemployment Capital Brief. The RBA's Expert Advisory Group had discussed the effect of AI on productivity at its meeting on June 3, 2026. In background, Bullock delivered a speech titled 'Technology and the Future of Central Banking at the RBA' to the 60th Shann Memorial Lecture in Perth in September 2025, and said then the RBA is not using AI to formulate or set monetary policy or any other policy. She had also said recent rate cuts were expected to support spending by households and businesses.

The broader context here is the collision between asset prices, potential output and fiscal arithmetic. Central bankers monitor equity valuations through wealth effects, cost of capital and credit conditions. A sharp repricing can tighten financial conditions even without a policy move. At the same time, productivity determines potential growth, neutral rates and debt sustainability. If AI lifts valuations without lifting measured output, policy faces looser financial conditions now and tighter supply constraints later.

Looking at what this means for policymakers, the intergenerational arithmetic is the pressure point. Long-run budgets built around 1.2% productivity growth leave limited margin if AI gains arrive as shorter hours rather than higher national income, or if datacentre buildouts add to demand before software diffusion reduces costs. Valuations do not set policy. Expectations embedded in those valuations shape consumption, investment and revenue forecasts that do.