More than robotaxis: the tech revolution and productivity

Adair Turner from the Institute for New Economic Thinking suggests the accumulation of wealth from this technology wave is being spent on low productivity activities. He points to expenditure on domestic help and entertainment and suggests that, in aggregate, measured productivity could actually decline.

"Even if technology doesn’t reduce the aggregate number of jobs, it certainly does disrupt the kinds of jobs”

Other research, cited by the RBA, suggests the uneven take-up of technology across firms may be behind slow productivity growth. Productivity growth is measured as an average across the economy. If the gap between leading and laggard firms is widening, average productivity may not increase despite some firms taking up new technology.

{CF_VIDEO}

This is fair enough. But we still don’t have an answer to an even more fundamental question: what will this revolution do to jobs? And will technology create more jobs than it destroys?

Whether there are more or less jobs created we will only know in time. What we can be confident about, however, is there will be very substantial change within the mass of jobs. Many roles will be destroyed; many new roles created.

In a recent follow-up to his must-read Homo Deus, Yuval Noah Harari delivers a powerful sense of the scale of changes we may see from artificial intelligence (AI). He argues the current wave of innovation is different because it is driven by technology that offers connectivity and updatability.

His critical insight: we shouldn’t compare human cognition with one machine; we should compare it with a network.

As the frontier of technology improves, the entire system can be updated to the new standard. “The AI revolution won’t be a single watershed event after which the jobs market will just settle into a new equilibrium. Rather, it will be a cascade of ever bigger disruptions,” Harari warns.

In 2018 the US Food and Drug Administration approved the first AI software that can make referrals without needing a specialist to interpret medical imagery. The software, IDx-DR, analyses images of the retina to diagnose diabetic retinopathy, the most common cause of vision loss among diabetics.

The software itself makes the recommendation about whether the patient needs more specialised care. A camera operator is the only human input required.

Investment only in software and a camera easily broadens the scope of diabetes services. It allows existing specialists to focus more on treatment and less on diagnosis. Which presumably means over time fewer highly trained medical specialists will be needed to deliver the same level of service.

Or consider autonomous vehicles. While there remain many issues to resolve, the progress seems undeniable. Tesla’s testing reveals much lower accident (or near miss) rates on autopilot - one every 3.34 million miles - than when the vehicle is under human control – one every 1.92 million miles.

That there has been some high profile robotcar disasters doesn’t change the aggregate picture. If you happen to be in metro Phoenix, Arizona, you can now summon a robotaxi from Google’s Waymo using an app on your phone.

So what happens to crash repairers, drivers, mechanics, call centre operators in the industry?

Historical experience is consistent with the idea that even if technology doesn’t reduce the aggregate number of jobs, it certainly does disrupt the kinds of jobs. And lives.

A recent report from the Pathways to Prosperity Commission in the UK which considers issues of technology and inclusive development suggests that during the Industrial Revolution working-class labourers experienced a decline in living standards for the first 60 years of this period - while the income of the top 5 per cent more than doubled.

Consider a specific example of the introduction of washing machines into households: in 1910 there were 500,000 domestic laundry staff employed in the US. Within 30 years their number had declined by nearly 90%.

Research from Alpha Beta looks at the impact of technological change on the labour market from two perspectives – the change in jobs across the economy and the change of tasks within specific jobs. Their key conclusion is that these two types of change are negatively correlated. In other words, the more the tasks within a job change to adapt to new technology, the less chance the job itself becomes obsolete.

Take the lesson from the predicament of domestic laundry staff a century ago. The best answer was to embrace the washing machine and make it integral to your working life. Trying to wash clothes more quickly by hand would not have worked.

The absence of a measured increase in productivity shouldn’t distract us from the profound impact technology is having on the way economies function.

We can’t accurately forecast if enough jobs will be created to replace those destroyed. But jobs will change and jobs will be challenged; that much is clear.

Richard Yetsenga is Chief Economist at ANZ