INSIGHTS: AI and the Complexity Trap. Lessons from Ancient Rome for Australian Savers.


Key Takeaways

  • Cheap imported goods have masked the rising cost of everyday services.
  • AI may be creating new paperwork faster than it saves time.
  • The first gains from AI may flow to businesses rather than households.
  • Ancient Rome met rising complexity by debasing its silver coinage.
  • Australia's past productivity booms followed periods of simplification.

Alex Hutchinson joins host Chris Tipper on Ainslie Insights to argue that artificial intelligence is more likely to add complexity to the Australian economy than to rescue its productivity. The discussion runs from William Baumol's "cost disease" to Joseph Tainter's study of ancient Rome and what both mean for Australian savers.

Cheap goods have hidden the rising cost of services

Alex Hutchinson argues that three decades of cheap manufactured goods, much of it from China, have disguised how fast services have become more expensive in Australia. She calls it the "China disguise": falling goods prices held inflation down while service costs climbed.

The China disguise rests on economist William Baumol's cost disease. Factories keep getting more productive, so goods get cheaper, while teaching or nursing still depend on a person's time and wages must keep pace (Yale University Press).

Australian price data shows the split between goods and services. Analysis of ABS consumer price data reported by the ABC in July 2019 found secondary education costs rose 203% between 2000 and 2019, while audio-visual and computing prices fell 89% (ABC News, 29 July 2019).

Why do Australians feel poorer when goods keep getting cheaper?

Chris Tipper framed the problem through the television. A household can own a far bigger screen than its parents did and still struggle to afford healthcare or education.

Hutchinson's example is people with a bad back delaying a visit to the physio. In her reading of history, the gains of the 1920s and 1930s came from appliances such as the washing machine freeing household labour, and nothing comparable exists today short of robots.

The counter-view comes from Baumol himself, who argued that because productivity elsewhere keeps lifting overall incomes, societies can still afford rising service costs (Yale University Press). The ABC analysis also notes that much of the fall in electronics prices reflects ABS quality adjustment rather than cheaper price tags.

AI may be adding work before it saves any

Hutchinson's most contrarian claim is that AI is currently increasing the amount of work in the economy, not reducing it. Every AI-drafted claim or contract mark-up still needs someone to answer it.

The Fair Work Commission is Hutchinson's lead example, and the published record supports the broad picture:

Fair Work Commission measure

Figure

Source

Growth in total workload over three years to 2025-26

More than 70%

HCAMag, 23 August 2026

Applicants who said they used AI (preliminary study)

40%

Clayton Utz, June 2026

Total lodgements, 2025-26

53,666, up 22%

HCAMag, 28 October 2026

AI disclosure rules for parties

In force from 20 October 2026

HCAMag, 23 August 2026

Commission President Justice Adam Hatcher attributes the growth principally to applicants' use of generative AI rather than to labour market change (HCAMag, 23 August 2026).

How is AI slowing down business deals?

Hutchinson described a private equity acquisition delayed because a junior lawyer used AI to contest the contract clause by clause. "Using AI to pull apart a contract is great. Give me the headlines and tell me what I need to look out for. Not every clause," she said.

A tradesperson Hutchinson spoke with now receives AI-written emails telling him how to do jobs he already knows how to do. "So the amount of work this is adding is astronomical right now," Hutchinson said.

Tipper added that using AI this way is rational: each party protects its own interests, so the work grows even when the outcome barely changes.

Is AI already cutting Australian jobs?

Some job effects are visible. Bunnings announced on 17 November 2026 that around 350 Australian technology roles, about 230 permanent positions and 120 contractor engagements, would go under an outsourcing deal with work shifting to teams in India and the Philippines (Inside Retail, 17 November 2026). Hutchinson was careful to say she did not know whether those cuts were driven by AI or by offshoring.

AI's early gains may go to businesses, not households

Hutchinson expects the first returns from AI to accrue to firms rather than to individuals. "AI will return to businesses, not to individuals," she said, adding the qualifier "at this point".

Where AI's gains land matters because Australia's productivity record is weak. The Productivity Commission says labour productivity growth averaged 0.4% a year over the last decade, the lowest in 60 years (Productivity Commission).

Hutchinson pointed to Robert Solow's 1987 observation that "you can see the computer age everywhere but in the productivity statistics", first published in the New York Times Book Review on 12 July 1987 (Stand-Up Economist). Computer-era gains did show up in the 1990s, she noted, but AI's gains may not arrive in time to ease inflation, so in her view governments should reduce spending.

A more optimistic view sits with the Productivity Commission. Its August 2025 interim report estimated AI could lift labour productivity by about 4.3% over a decade, adding around A$116 billion to GDP, while noting debate about the size of the effect (HCAMag, 8 August 2025).

Ancient Rome shows how complexity can end in debasement

Hutchinson draws on anthropologist Joseph Tainter's argument that each new layer of social complexity returns less than the one before. Tainter set out the thesis in The Collapse of Complex Societies, first published by Cambridge University Press in 1988 (Cambridge University Press).

Hutchinson's illustration runs in layers: an army pays for itself in security, a city police force still adds value, but officials weighing market vegetables add cost without much return.

Rome's response to mounting costs was to stretch the coinage. The denarius held about 95-98% silver under Augustus and only about 50% by the reign of Septimius Severus (193-211 CE) (World History Encyclopedia). Hutchinson sees a parallel in modern governments carrying rising debt and interest costs, and argues AI could add complexity in the short term without reform.

Rome's decline has many competing explanations. Tainter himself reviewed more than 2,000 years of them (Cambridge University Press), and modern central banks, including the RBA, operate under inflation mandates Roman mints never had.

Readers following how gold and silver respond to currency debasement can track live prices on Ainslie Bullion's charts and prices page.

Simplification has historically preceded Australian booms

Hutchinson's proposed remedy is simplification, and she argues Australia has done it before. In her view, the Hawke, Keating and Howard eras simplified the economy and were followed by investment and growth.

Hutchinson wants the same approach extended beyond tax, so that teachers can teach and doctors can treat without layers of reporting.

The Productivity Commission has flagged the compliance load too: boards spent 55% of their time on compliance in 2025, up from 24% a decade earlier (Productivity Commission, August 2026).

The harder question is consensus. Tipper asked how governments could agree on AI rules when models already sit outside regulation, and many existing rules serve safety and consumer protection. Hutchinson still expects technology to pay off eventually: "AI will help, robots will help, just like the washing machine helped," she said, while warning that the current phase is adding layers.

What it means for investors

Hutchinson's argument is about timing. If AI's gains arrive later than governments hope, Australian savers may face a longer stretch of rising service costs and public debt.

Tipper noted that Ainslie Insights often returns to gold and silver as a long-run response to currency debasement, calling them "certainly part of the answer". Ainslie Bullion offers physical gold and silver with vault storage, and Gold Silver Standard offers tokenised precious metals.

For a theme that may play out over years rather than months, Ainslie Saver is built for steady accumulation over time. Further Ainslie Insights discussions sit in Ainslie Bullion's News and Resources section.

About Alex Hutchinson

Alex Hutchinson joins Ainslie Insights from Siltech PMR.

Episode: "AI and the Complexity Trap: Lessons from Ancient Rome for Australian Savers", an Ainslie Insights episode.

This article is general information only and does not constitute financial advice. Past performance is not indicative of future results. Always conduct your own research or consult a licensed financial adviser before making investment decisions.