“Artificial intelligence is almost going to be like a utility, like electricity. You wouldn’t ask what areas of the economy or industries are going to be impacted by electricity – the answer is all of them, everywhere.”
At the current rate of developments in robotics and AI, it is estimated that by 2019, up to 25% of tasks across every job category face automation.
This raises a spectrum of questions. What form is automation likely to take? And how can people, organisations and governments prepare or respond?
Tackling these issues in this interview from PwC Consulting are Martin Ford (New York Times-bestselling author of Rise of the Robots: Technology and the Threat of a Jobless Future) and Managing Partner of PwC’s People Business, Jon Williams.
The first wave of the automation megatrend lies in the replication of routine or physical tasks on factory production lines. The current capabilities of robotic process automation mean that back-office tasks can now be performed by software. The “fourth wave” of machine learning and artificial intelligence, says Williams, will see robots perform complex cognitive functions such decision making in unusual or new situations.
Automation’s gradual advancement has raised questions over its ability to neatly replace existing roles or supplant the value of human interaction. In this video, Williams cites the example of a fully automated food court, such as the one at Singapore’s Changi Airport: it may be a great fit for international airports, but would customers want the same type of speedy, human-free experience in, say, a pub?
Organisations may consider a number of implementation strategies, such as restructuring partially automated roles or funnelling new resources into customer service or artistic or experiential endeavours – areas where the human touch is critical.
Governments, meanwhile, may need to revise the way they devise tax systems and the education sector to prepare for this new tomorrow. Other considerations include the way unions might respond to a workforce comprised of robots.
Clearly, the accelerating pace of change ignites discussion across a range of fundamental issues. What do you think the future could look like?
Watch the discussion on the rise of automation, or read the full interview below:
Neil Plumridge: Welcome to today’s webcast, Rise of the Robot: Is your organisation ready? I’m Neil Plumridge, PwC Australia’s Managing Partner for Consulting. Joining me today are two great speakers. We’ve got Martin Ford, author of Rise of the Robots – I’m looking forward to hearing Martin talk today – and PwC’s Global People and Organisation leader, Jon Williams. Welcome Martin and Jon.
Jon, as you know I’m very passionate about Australia’s place in the world. We’ve got an abundance of resources above the ground in people and we’ve also got a lot of resources under the ground in minerals. But I’m really excited to understand how we can deploy those human capital resources we have. But recently, a number of our colleagues wrote a report saying 44% of jobs over the next 20 years in Australia – that’s five million jobs – may disappear or be disrupted significantly through automation and robotics. I’m interested in your thoughts on this: is that something similar to what we’re seeing perhaps 20-30 years ago in the 1970s and 1980s, or is it a different phenomenon?
Jon Williams: Thanks Neil. So, it would be glib to say that this is the white collar version of the blue collar event that happened in the seventies. It’s an easy statement to say we’re simply seeing in offices what we saw in factories.
I actually think – and Martin may characterise this differently – of four waves of what happened and what’s going to happen. So the first wave was very clearly in the seventies and eighties, in the factories and controlled or regulated environments built to have standard tasks carried out by humans. Many of those were replaced by robots to increasing degrees of precision. So we went from gross motor tasks to fairly precise tasks over time in those environments. That’s the first wave and that’s kind of what you’re referring to.
The second wave was the shift into unregulated or uncontrolled environments. Self-driving trucks in mining sites, robot-driven gantries on the docks, in different environments outdoors that were not completely controlled. That wave is still happening and self-driving cars are probably the end of that wave.
The third wave is the one I think we’re mostly talking about, which is taking that concept into offices. How do we start to do routine standard office-based white collar tasks through robots, computerisation and machine learning? In particular that relates to the professions: accountants – people in the firm that we work in – tax professionals, lawyers, increasingly doctors from a diagnostic perspective. How do you take those routine, often binary decision-making situations and convert them into a task done by a robot or computer? That’s the wave that’s about to begin and that software is already in place. We know that robots or computers make better diagnostic decisions than doctors in many instances and make fewer errors.
In the fourth wave we’re going into non-routine white collar work. Decision making, responding to unusual situations, responding to new situations. That’s the wave we haven’t quite got to yet and that’s the one that will come in 10 or 20 years. Exactly what that looks like, we probably don’t know yet. Maybe Martin does. I’m not sure we quite know what that’s going to look like yet.
Neil: So you see the change happening at the same time across all industries? Here in Australia we’ve got the financial services, healthcare, a lot of government sectors…
Jon: I think that’s fascinating. This isn’t the first time – and the reason I’m relatively optimistic about what’s going to happen – this isn’t the first time this sort of thing has happened. A similar thing happened in the agrarian revolution. You go back hundreds of years and 99% of people worked on the land and they couldn’t possibly envisage what you could do if you weren’t working on the land. There were jobs they couldn’t envision at that time. The industrial revolution did the same thing.
When you think about each of those, they didn’t happen all of a sudden, at once, in every place on earth and in every industry. They happened in bits and pieces at different times in different locations around the world at different paces. It’s not a switch that’s going to be turned and everything’s going to be different; in different industries and geographies it’s going to happen at a different pace. It’ll reflect the nature of the roles that are performed, so the extent to which we can use computers and robots to replace them or augment them. It’ll affect capital investment, it’ll reflect risk aversion, willingness to change, the human decisions that will either trigger this or not trigger this happening at a certain speed.
You know what? We have choice. We do actually have choice in this space. I was overseas recently and read about Singapore introducing their first fully automated food court at Changi Airport. A food court where you go to the machine, tap in what you want to buy, put your credit card up against it, a machine makes the food, you go to a little hatch, it pops out and you get it and eat it. Four percent of Singapore’s population works in food production, but it’s only 1% of GDP, so they’re trying to balance that up a little bit.
At Changi airport I’d probably be happy to engage in that. Because I’m travelling, I’m tired, I don’t want to have to engage someone that might not speak the same language as me, I might make bad choices. There’s a whole bunch of stuff where that’ll be a good situation. If you go to Satay Street in Singapore, I absolutely want a human being to –
Neil: The experience…
Jon: Absolutely, so that’s a choice. We could have had fully automated pubs ten years ago. You can make a beer dispensing machine. We don’t want that. We’re actually choosing to have a human to interact with. So there are human choices in this, it’s not just an automatic process that will happen. It will happen at different paces in different places.
Neil: Fascinating. Martin, when you released your New York Times bestselling book there was quite a lot of criticism calling it science fiction and the like. As you’ve talked around the world, there are an increasing number of conversations going, ‘well this is no longer science fiction, this is real now’. Are you seeing that in the different places you go, that people are now thinking of this not as something that may happen but something that definitely will?
Martin: I think that people are taking it a lot more seriously. It’s an idea that’s gaining traction.
It must be said that there’s still not a consensus by any means. You could easily find economists – probably brilliant economists with Nobel Prizes – who would tell you that this is all silly and not something we should worry about because it’s happened so many times in the past. It goes back a minimum of 200 years to the Luddite revolts and since then this fear has come up again and again.
There are some people who seem to have this belief that there’s some kind of economic law that says people will always remain relevant and we’ll always have jobs. I don’t really see things that way. What I find is that, especially among technical people who are really working on these technologies like AI, there’s a kind of emerging realisation that this is going to be a disruption.
What it really comes down to is that the machines are taking on cognitive capability. They’re beginning, in a limited way, to think. That’s something that really is quite different to anything that’s happened before. The other thing is that it’s now a general purpose technology, something that will scale across everything. Again, that’s quite different to what we saw in agriculture, for example, where you had tractors and combines that were enormously disruptive but very specific to that one sector of the economy. They didn’t invade the other areas.
Information technology and AI is almost going to be like a utility, like electricity. You wouldn’t ask what areas of the economy or industries are going to be impacted by electricity – the answer is all of them, everywhere. The same is going to be true of AI. I do think we are looking at a bigger disruption this time.
Neil: Have you seen any particular country or any types of companies ahead of the curve and adopting this as first movers and others that are lagging? Or is everyone around the world adopting this at the same pace?
Martin: Silicon Valley is certainly leading the actual technology and companies like Google and Amazon and Microsoft are especially involved in research into deep learning, which is the driving force behind this. But it’s going to disseminate very rapidly, there are very few barriers for this technology to invade just about every industry. I expect it to really come online across the board. The more interesting question is: are there countries or political systems that are really beginning to think about the implications of this and wonder, ‘what shall we do about it?’
I’ve gone around the world and talked to some governments – mostly in Europe. The Netherlands, for example, seems very keen on thinking about this. Singapore has also been mentioned a number of times. I’m going to Austria later this week to talk to people in their ministry. So there are many governments, including Australia, that are at least thinking about this. I even went to an event in the White House under President Obama and had a discussion about this. I’m not sure what’s going to happen in the United States now over the next four years.
Neil: So what role do you think governments have? There’s always this debate about how the government has a role to pick winners or not – about big government versus small government. In this particular spot, with the technology moving so fast, do you see governments as observers of the technology or participants in funding its evolution such as backing startups?
Martin: Certainly government has a big role in driving research in the technology. But the role of government that particularly concerns me going forward is that I believe this has the potential to make inequality much worse. A lot of people are likely to be left behind, and it’s going to be those people that are best equipped to do things that are fundamentally routine and predictable. That’s probably a pretty good share of the population. People have always had that kind of a job and are probably best equipped to do that kind of a job. Those people are going to find it harder and harder.
Ultimately it’s going to be the role of government to step in and make sure inequality doesn’t get too extreme, people have access to some sort of fundamental safety net so they don’t fall too far. In an advanced country like Australia and the US we obviously hope that a huge fraction of the population is not going to fall below a certain level and begin to lose hope and faith in the system. Unfortunately we’re seeing some evidence of that already. That’s what I’m mostly concerned about. I think there’s absolutely an important role for government and policymakers in that arena.
Jon: I think government does need to think about how we pick up that portion of the population that gets left behind. I think Martin and I are in agreement that there’s going to be a period of dislocation of many people as this happens and those jobs disappear. Because when jobs go, they tend to go in swathes and are replaced in smaller increments. That’s just the way the world works. You don’t suddenly invent a thousand new jobs but you can make a thousand people redundant quite quickly. I think we disagree on whether or not it will work out in the long run. I take a more optimistic view.
Even in the short term, governments have got to grapple with this whole notion of a living wage. Do we provide a living wage to everyone, and if we do, how do we construct a tax system that pays for that? If you think about the current system, it taxes both corporates and individuals who create value for the organisation as they go. Do we start taxing the robots instead of the people? Or do we change the corporate tax system to pick up enough money to pay for those who’ve been left behind?
Governments also need to think about how it responds to unions. The union movement, which is still relatively strong and powerful in some countries, has a huge role to play in this. You can’t have a union for robots, but you can have a union for people. I was in the US a couple of weeks ago. The Teamsters union is the biggest union, representing people who drive trucks. Well, think of one of the most easily replaced jobs in the next five or ten years – it will be truck drivers. So how the union responds to that and how the government interacts with that will be fascinating.
My third point is how there’s been a huge wave of offshoring of both manufacturing and service delivery. Manufacturing may well become local again, rather than long distance, because we’ll be able to 3D print objects.
There’s a fascinating story of a plumber. These days, a plumber goes to your house and find the part that needs to be replaced under the sink. They go to the hardware store and buys the part and comes back and fits it. In the future, they’ll go their van and print the part and go fit it. So you get rid of logistics and you get rid of manufacturing overseas. And then many of the jobs that have been offshored are jobs that are easily replaced because they’re relatively routine and mundane. These can be easily replaced in a virtual way by robots. There’s a government interplay in that as well, how different countries collaborate to make that whole system work. I agree that government has a huge role to play in moderating the system.
Neil: Interesting stuff. We could keep going, but I’ll be keen to hear from the audience. The first question has come through and I might throw this one to Martin: what should people in the workforce do now – or even parents with children that are about to enter the workforce – do to prepare for the next ten years with a high degree of automation and robots?
Martin: What we definitely agree on is how work that is routine and predictable is going to be very susceptible. If you’re going to school, the last thing you want to be doing is studying so you can do a routine job. In the extreme, that can even be something like a radiologist – a doctor that interprets medical images. It takes an incredible amount of training to get to that point, and yet I think it’s a job that will someday be done exclusively by computers.
It’s not necessarily about how much education you get, it’s about the nature of what you’re training for. I think for the foreseeable future your best bet is to do something creative, where you’re generating something new as opposed to manipulating information in some routine way.
Or something that really involves deep interaction with other people, building relationships with people. Maybe something like nursing or being a doctor that works with patients, or in a business environment building deep relationships with clients. Those areas are going to be less susceptible at least over the short term – the next ten or 20 years. Beyond that, nothing’s off the table.
Jon: I largely agree. My youngest son just finished his high school exams. He’s been asking me this question most days for the past six months. He’s going into an industrial design course at university. That’s OK, it’s a bit creative, there’s a human requirement in there. I think the shift is going to be away from the notion of going to school, going to university, doing the same thing for 40 years, then having a rest and dying. This is still our current model for most people. You can’t possibly spend three or four years learning things that are going to be technically relevant for your entire career in the way that many of us did. You could learn to be an accountant and apply that for 30 years. That’s just not going to work.
We need to shift to a model where you learn a bit of stuff, you apply it, you know that might not last very long, and in five years’ time you learn something else. The model of lifelong learning is where we need to shift to, rather than expecting to learn to be a lawyer or accountant or tax professional up front then applying that forever.
I think we need to focus more on the skills of empathy, intuition and creativity. How do we foster those? There are many stories about how young children have those skills and then we beat it out of them with an education system still based too much on rote learning. So how do we develop those skills? I feel very strongly that we need to be focused on giving younger people in our society much greater resilience in themselves to be able to cope with what’s coming.
Neil: Some of these skills, are people coachable immediately? In the workforce today, 75% of Australia is made up of service-based jobs and industries. Five or six million people are reading Martin’s book and going, ‘well what do I do now?’
What would be your advice to those in those jobs that are in danger of either being automated or disrupted? What can they do now?
Jon: There’s a great saying, which is: Don’t rage against the machine, race with it. I think that’s the advice that everyone’s got to take. Accept that this is going to happen, that this change is going to come and figure out how you can apply yourself to get the most out of that situation.
We often say the spreadsheet didn’t kill accounting. It might have killed bookkeeping, but it allowed accountants to do something much better and more interesting and give advice and counsel rather than spend their time recording data. How can you use that analogy in your job? To be augmented by robots and computers that can manipulate data much better than humans, but a human can then use that data in a unique way to explain it to someone else and allow them to make decisions. So it’s about that acceptance of the change, awareness that it’s going to happen, and making the best of the situation.
Neil: We’ve got another question. Many people also currently gain their self-worth, self actualisation and personal satisfaction from their paid employment and – let’s face it – we measure each other by our jobs as well. To both of you, what do you think each of us and society as a whole needs to do to adjust emotionally and culturally to a post-work world?
Martin: It’s a challenge. Right now in the world that we live in, jobs are kind of a packaged deal. They come with an income, which is of course very important. They also come with this combination of something to do to occupy your time, and your sense of accomplishment and self-worth and so forth. Many people do derive a great deal from that. Not everyone; some people have jobs that they despise – we should remember that, too.
Clearly, as human beings we need both of those things. In the future we’re going to need an income, and we’re also going to need this sense that we’re contributing something that has purpose. Those don’t necessarily have to come from the same place, we can look at a future where perhaps those things decouple. They’re both there but provided by different sources: a guaranteed income or a universal basic income providing the lion’s share of our income. But then we find other things to do. There are a number of things that people do that fulfill them. You can work in the community as a volunteer. You can edit Wikipedia. I think there are many examples where you can solve that problem.
Jon: Yes it’s a really difficult question. If you look at what we’ve just seen in the US election, an awful lot of the people there are disaffected voters from a wave of dislocation that happened 20 years ago. There are stories of factory towns where the factory closes down and hasn’t been replaced. We haven’t as a society found something better for those people to do. I’m optimistic that in the long term there won’t be – this is where I disagree with Martin – a post-work world and that people will find or choose things to do.
This is a strange analogy, but I’m always fascinated by the TV show Masterchef. Ten years ago, food was a pork chop on a plate with a couple of potatoes. Now we’re spending time carving carrots into beautiful shapes. We’re curating experiences for our fellow humans that you wouldn’t have called a job and you wouldn’t have even thought of spending time doing – because you didn’t have the time to do it. We will free up the time for those artistic endeavours creating a more perfect world or experience for each other. It might be called work or it might not.
If you think of the agricultural revolution, they would look at people doing jobs now and say, ‘That’s not work! Work is getting your hands dirty’. We’ve evolved our view of what work is. That’s kind of what Martin was saying, people might do things that we wouldn’t consider work now, but they give purpose and meaning to your life and maybe that’s work.
Martin: That’s right. But the question is, are you going to be paid enough and will there still be a market for that? There may not be, and I think that’s an argument for having perhaps some kind of universal basic income in combination with these opportunities to do things that might add incrementally to that income and also occupy your time. I think that there’s going to be a combination of solutions in the future.
Jon: People want to do stuff. We all know that we’re way more productive in real terms than we were 50 years ago, 100 years ago, but we haven’t stopped working as hard. We’re making the choice. There’s a huge drive towards wanting to find things to do, whether they are things that are currently called jobs or not. How we’re going to afford that is an open question – the Swiss voted on a living wage earlier in the year. It was voted down, but it got approximately 30% of the vote in favour. This is a very real concept that might happen.
Neil: Another question – there’s a debate on whether there will be partial jobs or activities automated or full swathes of roles. What’s your view?
Jon: It’s mostly partial jobs. Few jobs are completely made up of tasks that can be automated, or at least automated at the same time. It might occur in sequence.
We have a choice though: do we simply take one person and have them do the five bits left over from the five jobs that were replaced, or do we say: ‘Let’s keep the five people and we’ll allow them to do something else in an augmented way because some of their tasks have been done by computer’. One is an efficiency play while the other is an increased delivery or customer service play. You can respond to the same situation in either way.
It’s much less whole jobs going in this wave than it is parts of jobs going and us having to find a way to restructure tasks.
Martin: I would tend to agree with that. For the most part, the impact is more holistic on the organisation. It’s not a one-to-one correspondence of a robot to a job. Technology redefines the boundaries between jobs and shifts the work in entirely new ways. Again, that certainly will result in consolidation in many cases. If you automate half of what two people are doing, then in many cases that will still result in the loss of one job. There are exceptions to that. If you think of self-driving cars, that’s a direct substitute for a taxi driver. But for the most part it’s not a one-to-one correspondence between a robot and a worker.
Neil: It’s been fascinating to talk to both of you. Martin Ford, New York Times bestseller of Rise of the Robots and Jon Williams, Global People and Organisation leader, thank you for joining me. And thank you for joining us on this webcast today.