Dispelling Dystopia: why many jobs will change rather than disappeaR

19 March 2018

Professor Hugh Durrant-Whyte, Director of the Centre for Translational Data, The University of Sydney


I’ve spent most of my life building robots and studying Artificial Intelligence (AI).
And I feel it's time to dispel some myths about AI, the future workforce and the rise of technology in our lives.

When I was first studying AI back in the 1980s, I read, what at the time was probably one of the only books on AI. The book was by Patrick Winston, the first director of the AI lab at MIT.

In it, he defined AI as anything that we currently can’t do. If we know how to do it, it's an algorithm.

In the early days, we thought of AI as based on models. The paradigm shift has been that we no longer start with models, we start with data. And from that data, we create models to make predictions.

But there are lot of divergent views about AI and its role in the world. One view is that AI is the future of all human kind — whoever is the leader in this sphere will become the ruler of the world.

Another view is that held by Andrew Ng, who is well-known in AI. He says he worries about AI super intelligence in the same way that he worries about overpopulation on Mars. Yes, it is a concern, but it’s such a long way away, it’s not actually front of mind.

What we see now is AI changing values and disrupting business economics.

And we see it everywhere.

Australia has more autonomous vehicles working than any other country in the world. In fact, probably more than all countries put together.

There are now over 1000 robots in the Pilbara. There are autonomous trucks, autonomous trains, autonomous drill rigs — all remotely operated from one place in Perth. Rio Tinto is now putting these in a lot of their mines right across the globe.

Similarly, Port Botany, Port of Brisbane and increasingly Port of Melbourne are completely automated. We are replacing what used to be drivers and straddle carriers with autonomous vehicles and autonomous components. In agriculture, there’s also an increasing use of robotics in remote areas.

Australia leads the world in this kind of robotics because so many of our jobs are remote and automatable.

This is happening today and putting people out of work.

Job replacement has happened since time immemorial. But increasingly, there is a threat of elimination of entire areas of work and this has profound consequences, including rising inequality and societal change.

We are seeing a big change in what people do and in the way industries adopt robotics. At the same time, we are trying to understand what the impact may be.

Often, I see newspapers report that 40 per cent of jobs are going to disappear because of automation. In fact, the data shows that 40 per cent of jobs will change.

There are going to be different types of jobs that go and different types of jobs that remain. In particular, those jobs which require either engagement with people — aged care or waitressing for instance — or higher creative and technological skills will stay because they’re very hard to automate.

But a lot of the jobs in the middle — everything from an accountant, to even a lawyer and a doctor – they are easy to automate.

There’s going to be this hollowing out of industries and a polarisation in the job market. This is already happening in the US and Europe.

It’s not just a function of manual labour or non-manual labour. It’s the fact that certain jobs are just simply automatable, because they’re predictable and they involve analytics.

Jobs are being geographically polarised as well. Based on current industry, there are going to be fewer jobs in remote communities — in mines, agriculture and other areas.

Those communities will see depopulation, while places like Sydney will continue to attract those in the creative industries who want to work together.

Remember there were predictions that we were going to telecommute from the beach? We’re just not doing that. We’re getting together in these little hubs and if you’re outside of those hubs, that’s where we see jobs depleting.

But some visions of AI that enables robots to "take over" are just unfounded. Believe me, I produce them and I’m programming them, and the robots are not going to take over any time soon.

Jobs that are predictable, repeatable or analytic are going to be lost. And the jobs that require technical, creative and social skills will grow.

We will see a drive in productivity through increased automation, as has occurred in the mining applications. Equally, on the geographic side you’re going to see an urbanisation of jobs and skills.

This polarisation will drive inequality and the backlash that we’re seeing across the globe.

That is a challenge that we will have to face, particularly in Australia.

We’re not heading towards a dystopia in robotics and automation. We are heading towards big societal changes, that both universities and the community will need to deal with.

This article is based on elements of a speech delivered by Professor Hugh Durrant-Whyte, Director of the Centre for Translational Data, The University of Sydney at the 2018 Universities Australia's Higher Education Conference.