Australia is investing heavily in artificial intelligence, data centres and other digital technologies. Businesses are experimenting with AI tools, global technology companies are expanding their infrastructure and demand for computing power is growing.
But there is a problem economists are watching closely.
The productivity boost that many people expected from AI has not yet clearly appeared in Australia’s economy.
Recent comments from Reserve Bank of Australia Governor Michele Bullock highlight the issue. She said there are still very few signs that AI is influencing the supply side of the Australian economy in a major way. At the same time, the AI investment boom is already adding to demand.
That creates an unusual situation: Australia is spending more on AI before it is clear how much extra economic output that investment will eventually produce.
Australia is adopting AI, but much of it is still early
AI use among Australian businesses has increased rapidly, although adoption remains far from universal.
The Australian Bureau of Statistics reported that 12 per cent of businesses used AI during the 2024–25 financial year, compared with just 1 per cent in 2022–23.
The RBA’s own survey of 105 medium and large businesses found that around two-thirds had adopted AI in some form. But the depth of adoption varied considerably.
For many companies, AI was still being used for relatively simple tasks such as summarising emails, conducting research or assisting employees. Only a smaller group had begun integrating AI into important business processes.
That distinction matters.
Downloading an AI assistant and using it occasionally does not necessarily transform how a company produces goods or delivers services.
The productivity numbers have not caught up yet
Australia’s broader productivity figures show why economists remain cautious.
The Productivity Commission reported in September 2026 that overall labour productivity was flat in the June quarter after falling 0.6 per cent in the March quarter. Over the year to June, output increased by 2.1 per cent while hours worked increased by 2.3 per cent, resulting in a 0.2 per cent decline in labour productivity.
Market-sector labour productivity increased 0.2 per cent in the June quarter but remained 0.1 per cent below its level a year earlier.
The RBA has also described Australia’s recent productivity performance as weak. Its August assessment said non-farm labour productivity increased by only 0.1 per cent over the year to the March quarter and fell 0.6 per cent during that quarter.
So far, the arrival of AI has not produced an obvious economy-wide productivity jump.
That does not prove AI will fail to deliver one. It means the transformation is still too early to see clearly in national statistics.
Why the gains may take years to appear
New technology does not automatically make an entire business more productive.
Companies often need to change their processes, train employees, reorganise teams and work out where AI actually provides value.
The RBA found that Australian firms expect AI and other technology investments to improve productivity over time, but many also said they need changes to workflows, workplace culture and employee skills before those benefits can be fully realised.
This can create a delay between investment and measurable productivity.
A company might spend money on AI this year but spend another year changing its systems and training workers. The productivity benefit could appear later.
There can even be an initial cost.
Businesses may temporarily become less efficient while employees learn new systems and existing processes are redesigned. The RBA has noted evidence from overseas suggesting that some firms can experience a short-term productivity decline before longer-term benefits emerge.
The AI boom is already boosting demand
There is another side to the story.
AI may eventually increase the economy’s ability to produce goods and services, but building the infrastructure needed for AI also requires huge amounts of investment.
Australia is seeing strong investment in data centres and related infrastructure. The RBA said business investment in the March quarter was much stronger than expected, driven by data-centre investment, while expectations for future investment have also increased.
This spending can support construction, technology imports and other economic activity.
But investment itself is not the same thing as productivity.
If companies spend billions building data centres but the wider economy does not yet produce significantly more output with the same amount of labour and capital, the productivity numbers may remain weak.
This is part of the awkward timing that the RBA is watching.
The real test will be how Australian businesses use AI
The next stage of the AI story is likely to be less about how many companies have access to AI and more about what they actually do with it.
There is a major difference between using AI to write an email faster and redesigning an entire business process so that the same number of workers can produce substantially more.
The RBA has found that many Australian businesses are still experimenting. Firms are interested in more advanced applications, including systems that can carry out tasks with greater independence, but practical adoption remains limited.
That means the biggest productivity gains, if they arrive, could come later as companies move from experimentation to deeper integration.
Skills will also matter.
Workers may need training to use AI effectively, while managers will have to decide which tasks should be automated and which still require human judgement.
Australia has more than an AI problem
It would also be misleading to suggest that AI alone will solve Australia’s productivity challenge.
Businesses surveyed by the RBA identified regulation, access to suitable workers and other structural issues as important barriers to improving productivity.
The Treasury has similarly highlighted the importance of technology diffusion — the process of getting existing technologies adopted and used effectively across a broad range of Australian businesses.
This is particularly important for Australia because the most advanced AI systems are often developed overseas.
The economic benefit depends partly on how quickly Australian companies can adopt those technologies and turn them into better products, lower costs or more efficient ways of working.
What happens next could matter more than today’s AI spending
The Australian AI boom is therefore entering a different phase.
The first question was whether businesses would adopt the technology.
The next question is whether they can turn that adoption into measurable economic gains.
There are reasons to expect productivity to improve over time. The Productivity Commission has estimated that AI could generate a significant productivity dividend over the coming decade, although it stresses that the size of the eventual benefit is uncertain.
But the latest Australian figures show that the payoff has not yet appeared clearly in economy-wide productivity data.
For the RBA, that distinction is important. Governor Michele Bullock said in September that AI remains a major potential source of productivity improvement, but there are currently few signs of a large supply-side effect. She also stressed that stronger productivity growth will be important for Australia’s economic performance over the longer term.
Australia’s AI story, therefore, may not be about whether the technology works.
It may be about how quickly Australian businesses can move from experimenting with AI to fundamentally changing the way they work.
Until that happens at scale, the country can have a very visible AI boom without seeing the productivity boom that many people are waiting for.