Australia’s Central-Bank Chief Questions Whether AI Is Lifting Productivity Yet

Reserve Bank of Australia Governor Michele Bullock has challenged claims that artificial intelligence is already lifting the country’s productivity. Speaking at a business event on September 22, she said there is not yet evidence of an economy-wide efficiency boost, even as governments and companies make large bets on AI.

Bullock pointed to research from South Korea suggesting employees using AI produced roughly the same output while working about 1.5 fewer hours each week. That result is not proof that AI cannot increase productivity; it illustrates that time savings and measured output are not automatically the same thing.

The implementation gap

New tools can require training, workflow redesign, data cleanup, and review processes before they change what an organization produces. Early investment can even add costs before measurable benefits appear. That makes it difficult to attribute national productivity trends to AI while adoption remains uneven and firms are still learning where the technology helps.

The distinction matters for public forecasts. If expected AI gains are built into long-term economic projections, policymakers need evidence about how quickly they arrive, which sectors benefit, and who captures the gains. A technology may create real value for particular tasks without yet changing aggregate productivity statistics.

Bullock’s comments are a reminder to separate the scale of investment from demonstrated results. The next test is not just whether businesses deploy AI, but whether they produce more or better work with the resources they use—and whether those improvements persist beyond pilot projects.

Sources: The Guardian’s report; Reuters on the RBA’s broader economic outlook.