The headline is a scenario, not a cheque

The modelling traces occupational productivity assumptions through industries, investment, employment, and the wider economy. Its base, low, and high cases show a range of possible gains if adoption raises output and those gains flow into additional capacity and demand.

That conditional structure should remain visible whenever the largest number is repeated. Real benefits will depend on implementation quality, complementary investment, competition, infrastructure availability, and whether firms use productivity to expand rather than only reduce labor costs.

Growth can rise while particular workers lose

Construction, wholesale and retail trade, and transport and warehousing benefit in the model as investment and household demand expand. Agriculture and mining move in the opposite direction because automation and efficiency reduce the number of workers required for a given level of output.

An aggregate employment gain is therefore not a transition plan. Governments and employers need location-specific training, portable credentials, income support, and practical routes into expanding sectors before displacement arrives. The most credible productivity agenda will publish realized outcomes by industry, region, wage level, and worker group rather than celebrating GDP alone.

Primary trail

Go to the source

Read the evidence behind this analysis. External links open in a new tab.

EY Australia — AI productivity gains could add up to $116 billion