Our workforce has been transformed, so what happens next is still up to us

Is artificial intelligence going to take chunks out of the Australian workforce? What are we mere humans going to do with ourselves when AI hits its stride?

These types of questions are being posed more frequently, with news of job cuts in Australia, and a tough labour market for new university graduates in the United States as uptake of AI is cited as a cause of fewer entry-level jobs.

In expert circles, there is heated debate as to whether the impacts of AI on the workforce – particularly in reducing the need for white collar professional jobs – are imminent and profound, or dramatically overstated. Much of this disagreement is entirely reasonable: history supports healthy scepticism about tech replacing (as opposed to transforming) jobs.

The more prosaic fact is that we have been in the midst of a quiet transformation of Australia’s workforce for decades: the shift of jobs away from agriculture, then away from manufacturing, to services. Four out of five Australian workers work in services, reflecting the welcome move from manual and routine work to non-routine and so-called “cognitive” work.

The economic literature documents two effects of new technology: displacing workers from some tasks (the displacement effect), but creating new tasks where humans have a comparative advantage in (the reinstatement effect). Sometimes the displacement effect is the larger one in a particular industry.

When that change is gradual, “labour markets” – that is, the decisions of employers and current and future employees – have time to adjust. If a workforce declines by 5 per cent a year, that gives 20 years – a generation – for people to move. Younger workers decide to switch, school-leavers look to other paths. Changes are never quite that linear, but the time horizon is a key factor in how much of a shock these changes are.

Despite generative AI’s rapid development, there are good reasons to think that it, alongside other forms of automation such as robotics, will also be gradual in its impact on workforce needs. It will take time for us to learn where and how to make most use of automation: for example, gen AI’s strengths in parsing large bodies of data, pattern matching, imitating and recreating text and images.

Early adopters will have successes, but also failures, and over time we will calibrate the role of humans and the tools that work alongside us. But we need to acknowledge the potential for the impact on parts of Australia’s workforce to be much more sudden. Recent work by the International Labor Organisation points particularly at roles such as data analysts, bookkeepers and other administrative workers being most at risk of being replaced by AI.

For an optimist, automation will enable skilled workers to do more: think of the potential use of technology by a health practitioner to reduce mistakes, expand the scope and bring down the cost of healthcare.