What does automation, machine learning and AI mean for the future of mining, agriculture, cities, the future of jobs and the risks of a polarised society? Sandra Peter recently sat down with a leading robotics scientist and the Chief Scientific Adviser at the UK Ministry of Defence, Hugh Durrant-Whyte, to explore the next 10-15 years.
Hugh Durrant-Whyte is also a Professor, ARC Federation Fellow and Director of the Centre for Translational Data Science at the University of Sydney. In May 2018, the NSW Government appointed Professor Durrant-Whyte the next NSW Chief Scientist & Engineer.
Show notes and links
Thomas Piketty’s book Capital in the Twenty-First Century
Centre for Translational Data Science at the University of Sydney
Sandra Peter Introduction: Automation, machine learning, and artificial intelligence mean for the future of mining, agriculture, cities, the future of jobs, and the risks of a polarised society.
Introduction: From the University of Sydney Business School, this is Sydney Business Insights the podcast that explores the future of business.
Sandra Peter: I'm Sandra Peter and I recently sat down with a leading robotic scientist and the Chief Scientific Adviser at the UK Ministry of Defence, Hugh Durrant-Whyte, to explore what will happen in the next 10 to 15 years.
Hugh Durrant-Whyte is also a leader in robotics and Professor and Director of the Centre for Translational Data Science at the University of Sydney and an ARC Federation fellow.
We want to have a look at the future of automation. Where is the future of automation in next 10-15 years?
Hugh Durrant-Whyte: I think we've definitely reached a tipping point where occupations like mining are not just increasingly automated but in the next 10 or 15 years likely to be fully automated. You won't see anybody in any of these remote mining or remote oil and gas situations. It will all be automated and it will all be operated and controlled in a very automated manner, you're already seeing big operation centres that look after these places and increasingly I think you might even see one operation centre that perhaps controls hundred mines right across the globe or from one place. So I think those sorts of things are not just there but they will definitely happen because there are economic drivers and technical drivers to make that happen. I think the other big areas are likely to be agriculture which I'm not so sure full automation is the right kind of word. You always sort of to some degree need a farmer or a manager or whatever else but it will be increasingly one person, very large farm operated, even herding, milking, and all these things are now fully automated or can be fully automated and I think in the next 10 to 15 years you will see them really turn into just single economic automation kinds of things.
I was in New Zealand and I heard a talk on artificial foods and I think there's a very strong possibility that things like beef and even dairy will disappear as agricultural outputs because I think they'll just be produced by bacteria which is essentially what they're doing in vats. There's been lots of articles recently and I suspect within the next 10 or 15 years that will happen and that's probably a good thing because the greenhouse gas effects, the issues with cruelty to animals and things like that will disappear. So I think agriculture is the kind of thing that is on the verge of enormous disruption and will be almost a zero employer in the future. And I think increasingly despite what everyone thinks I think will just be grown rather than bred in any kind of way. So I think that's important and I think that's critical for Australia because it's hard to think of any sustainable economy that would exist in the regional areas of Australia.
Once you start removing mining and you remove agriculture and things like that side, I think there's a real worry in my view about what will happen in regional Australia - it's already been depopulated to a very very large degree and I think that will just carry on.
Sandra: So in 10/15 years of geographic polarisation?
Hugh: Even more than it is now. Although I think the numbers are there already. I mean young people don't stay in the country. And it's increasingly hard to sustain populations that live a long way from urban regions.
Sandra: Because even if these farms are automated or the mines are automated then you could control them basically, off shore or even from a different country.
Hugh: Yes, I've heard it also said "so why are we controlling them from Perth when we could be controlling them from Bangalore?".
So I think there's an element that says that nothing really needs to happen in Australia other than the robots digging the stuff out. So I think that's dystopian. But equally it's just hard to see in those sorts of areas, I mean they're economically going to generate money but I just can't see them creating jobs and I think that's the fundamental issue with automation is at some point you don't create nearly as many jobs as you destroy. And I will say we've reached that point in my view. It's more true I think in Europe and the US, but in more qualified areas there's certainly less demand for graduates now than there used to be. And we're beginning to see that, in my view, in Australia as well.
Sandra: Are cities going to change a lot?
Hugh: I don't think robotics is going to actually change what goes on in cities very much at all because dealing with people is something that robots are not very good at and so remote rural kinds of occupations, yep, you can get robots to do things because you can keep people out of the way. But getting them to operate in towns I really think is much harder. There's been a lot of hype about self driving vehicles and I've probably built more self driving vehicles than most people and I will say that we're not going to have a self driving taxi in the middle of Sydney in my lifetime. It's just too hard. And there are an increasing number of reports of people in Google and Toyota and everywhere else agreeing with basically that kind of sentiment.
Sandra: Why is it hard?
Hugh: Because people are very hard to understand. Computers need to understand things if they're to do things with them. I think there are still issues with sensing, perception as we call it, the ability to make images, interpret images, and all sorts of things and just being able to model and understand the fact that things are not actually logical a lot of the time, right? So I don't think you're going to see that. I mean people have been trying to sell robots for offices for a long time as well. And they're still people try and do that. But the reality is we've not really seen any other than vacuum cleaners and even they don't work that well, right? And I think it's just hard to justify why you would want automation to be in that role. I'm going to come back to that I think in a bit about where the jobs are going to go.
So what are you going to see? I think the other thing that's counter-intuitive that you're definitely seeing is people are again congregating. We used to think that we would spend our time on the beach telecommuting and that's actually exactly not what's happening. We're finding areas where we find people who do things in common whether it's medical devices, or arts and entertainment, or whatever and there's a precinct set up, there's a hub, in the same way that Silicon Valley perhaps is or before that Birmingham in steel and Manchester in cloth and people congregate around there. They congregate to discuss, to trade ideas, to work together and that sort of thing. So I think what you're going to see in cities is much more specialisation and you have to be an important hub to really compete in any kind of way. So places like Sydney and Melbourne are probably above the bar in the sense of having enough hubs that you can see jobs creation and you can see new ideas coming out in Sydney fintech, net hubs and things like that I think that working. I think more problematic are places where there aren't that many hubs or unlikely to be. Adelaide hits me as somewhere that really is just struggling to be of any critical size to build an economic future without the defence subsidies essentially that are going on in shipbuilding and things like that. I mean other areas that worry me in places like western Sydney I just can't see any hubs forming in that kind of environment.
So I think 10 or 15 years from now you're going to see a very much more polarised geographically society. I think you're going to see fintech people in the city, going to see advertisers out at Bondi. You're going to see medtech people in North Sydney and you're going to see nobody anywhere else and that's probably culturally not a good thing but it's definitely what's happening. And historically it's really just to repeat but with the stronger gravity and magnet of things like social media and email make it all stick together.
Sandra: So where will the jobs be and where won't they be?
Hugh: The kinds of jobs that computers, so more broadly artificial intelligence and machine learning, rather than robotics are removing are those that are really around analytics. So things where people sit down and analyse things whether it's a bank statement, whether it's your credit record details, whether it's your shopping profiles those sorts of things are very easy to automate and they're to a large degree what I think it was the middle income kind of jobs. And the reality is it's those jobs that are actually going. It's those jobs we've got to realise are not the future.
Sandra: So in finance, in medical industries?
Hugh: Yes I think medicine because it's a bit of a closed shop would be a longer time. But certainly law, increasingly law graduates are struggling to get jobs because a lot of basic work, introductory work, when you start getting on in law is all being automated. Searching, you know building case histories, things like that. Lots of software products out now to do that. I think we're only just at the beginning of it though. I think one of the big things that's come along in the last two or three years is reliable speech understanding right. So I think a lot of those sorts of roles of understanding text and converting that into diagnosis or evidence or other sorts of things, the jobs that are used to must disappear. So I think whether they're in the right areas, you know this idea that of order 30 to 40 percent of jobs will change in the next 10 15 years is probably true. I think it's likely to happen a lot in the sort of more conventional jobs you see in the city around Martin Place, around banks, around standard business school outcomes if you like just because that kind of analysis is the sort of thing that can be automated. There will be lots of jobs that remain. And I think there's a positive element to that - jobs that are likely to remain are those I think that are creative and ideally creative when it comes to applying technology, to solving problems whether it's figuring out new markets, whether it's figuring out how you can sell new types of things or build new kinds of markets. I think that's likely to happen.
The other end of jobs are those that are very very engaged people because like I said robots and AI are not very good with people. So things like domestic help, health care, that sort of thing I think is going to be still quite hard to automate and I can't see many changes in that area and other things like waitressing. It's just too hard to flip a burger with a robot frankly. You may redesign the process so you don't have to have someone doing that sort of thing. But the reality is it's economically almost never worthwhile. So I think my worry there and many people have pointed this out is there's going to be this polarisation between jobs - people who can be creative and do stuff and people who basically serve others and the jobs that are in the middle of the ones that are disappearing. And it makes me a little concerned that there is this polarisation and again there's been a lot of work particularly in Europe around the fact that jobs are increasingly polarized particularly amongst graduates. And of course there's Piketty's famous book which is also talking about economic polarisation as well. So I think that's the real issue. And in Sydney, let me just try and be Sydney centric, I think Sydney follows the world by you know five to 10 years. There are already big signs that graduates are struggling to get jobs. What we will need to be teaching as a skill set at least going forward and what it isn’t certainly a lot more analysis.
Sandra: So how do we think about the skills that we need to teach them going forward and how do we move people from being inventors, which is what there was a lot of emphasis on, to being innovators in terms of business models and application of technology.
Hugh: So the inventing part is to some degree like what's going on now. People learning how to code, write programs, and stuff. And sure that's great and I'm really positive about it. The problem is it actually doesn't sell anything. What you really want is people who know how to apply that creativity to making something new to creating a good outcome and to do that to be honest you don't actually need to know how to code. You just need to know people who know how to code. And that's a very fundamentally different thing. We need to somehow, I know this is not easy and I'm not sort of advocating it in its simplest form, we need to kind of teach entrepreneurialism. Now it's about as woolly or vague as it can get you can't teach someone to be entrepreneurial. But we need to teach people to understand technology even if they're not technologists.
So you know taking someone, let's say in the business school will be a good example, and getting them to understand what computers do, what will quantum computing do, what will robotics do, what all these things do. And to think of markets out there and to understand that a market isn't a robot, a market is a new product or thing to which robots can be applied or computing to be applied or something else technologically could be applied. And so in some sense, in the business school perspective certainly, you want to be able to teach the application of technology to solving problem. And that for me is I guess the entrepreneurial part of it. It's worthwhile also looking at the low end again. So it's an interesting little observation that I guess maybe 10 years ago, maybe a bit longer, we used to clean our own houses, launder our own shirts, mow our own gardens etc. etc.. Now interestingly a lot of that is farmed out. You very often see people who just send all their shirts to the laundry, who would never have done that in the past - who have people come in and clean their houses, who have people who come in and mow their gardens and so on. So there is a growing service industry and there's no question about that. So I think it's wrong to say we're going to lose 40 percent of jobs. I just think they're going to be different. I think a lot of it is this kind of increased amount of service activity because really truthfully these polarised people who have the money to do it will employ more and more services to do what they used to do themselves. And I think that will certainly happen.
Sandra: Will there be a huge social cost to that?
Hugh: Yes I think that's absolutely true and this social polarisation I think is the biggest challenge that robotics and automation and so on will come because there will be a growing social divide in my view because automation is not happening blue collar upwards like it used to, it's happening in the white collar middle and I think that's critical. I also think things like the other opportunities. Let's pick something like MOOCs. I think one of the reasons that MOOCs perhaps have not been as successful is yes you can give a lecture to a million people but actually you need that personalised bet that applies it to you. And I think ironically if you like increasing automation well open the opportunity for individual people to have their stuff individualised by other people. So it becomes more personalised if you see what I mean in that process so I think some of these broader things, you're going to have people wanting to come in to create my online persona or my avatar or help me with my MOOC or whatever else it might be. So you kind of do get this global stuff but also lots of this local stuff. So I think there might be opportunities around that.
Sandra: What do you worry most about the future?
Hugh: Globalised IT companies I think are a big worry, an increasing worry. I'm sure everyone's read The Circle and things like that. What's scary about The Circle is it sounds so true.
I think our big thing in the future is going to have to be figuring out the whole data ownership, data privacy issue because I think actually most private companies have gone way beyond what any government would do, already.
Sandra: Is it too late?
Hugh: No I don't think it's too late. I think there are interesting technologies out there that have the opportunity to force change. For example, I mean as we are seeing in many cases, fully encrypted communications, fully encrypted engagement with the Internet. It's perfectly possible to run the Internet and do Google searches without people ever finding anything at all about you. If you were actually concerned about it.
Sandra: How do we get people to rethink ownership...?
Hugh: I think you already are I think it's just a slow movement. Thirty percent of people now use ad blockers, that's often kept quiet because it doesn't help a lot of people but if you went from 30 percent to 70 percent? How would you do online advertising? And then what's your business model if you're Google? The social polarisation is my biggest fear, no question about it. I think technology will come. I don't think it'll be as quick as people think it is. You only have to go back 10 or 15 years to realise that. Yes I know we've got smart phones and everything else. But the reality is I'm not sure we do that many things differently now than we did 10 or 15 years ago. And yes things will change. But the biggest change will be that one of disruption in a social context. Put it that way. I'm intrigued to see that in Europe the debate has started with this idea of a universal wage as an example. Not sure it's a good idea but nevertheless people are starting the debate. No one's even thinking close to that here and in the US the debate's been squashed. And I think the telltale figures in the US are not the unemployment rate it's the employment rate and the employment rate is now lower than it was in the Great Depression.
Sandra: Do we have other solutions other than guaranteed minimum wage?
Hugh: Not that I know but that's the subject for the business school right? Us engineers just automate things.
Sandra: So you get us in trouble, we get you out of trouble.
Hugh: That's right. Yeah I think that's by far the biggest worry, it makes people angry.
Sandra: What's the thing you're most excited about.
Hugh: The potential benefits that actually technology can bring to society and in a kind of exactly the opposite kind of sense right. I think about increased safety, I think about health, I think about understanding mental illness, I think about a whole range of possibilities of things that can happen. And they are happening, don't get me wrong, there's a whole cohort of people out there doing really globally beneficial things. There was a guy here who I like who is applying machine learning and AI to predicting future genocides for example and on a mission to basically try and prevent them before they happen. Now that's an application of AI that would otherwise put you out of work right. So I think there are lots and lots of very beneficial things that technology is providing and at an individual level I think it's a positive thing, I mean we've all benefited from just that growth of our social circle by having things like phones and Facebook. I think that sort of thing's great and positive. But I think there's always a flipside, right, with that they track you, they sell you things, they become monster companies who don't seem to understand the limits of that power.
Sandra: Will they still be around in 10-15 years?
Hugh: I hope not. Let me put it this way I think without going into too many details I think the technology is out there to limit the power of these companies by making data much more personalised and hidden. And for people to engage yes in a social media kind of manner. But without that data actually being gathered or utilized by third parties effectively disintermediation the disintermediators. So I think that'll happen. And increasingly you know I mean the Productivity Commission recommended that approach in a number of areas in their recent report and I can see that kind of technology coming to the fore. Having said that, these are not stupid companies right. If there is a technology out there that might disrupt them they tend to buy it. So I don't write them off that quickly but I think it will increasingly be the case that people will be able to control their future if they wished to.
Sandra: Thank you so much for joining us today.
Outro: You've been listening to Sydney Business Insights, the University of Sydney Business School podcast about the future of business. You can subscribe to our podcasts on iTunes, SoundCloud or wherever you get your podcasts and you can visit us at sbi.sydney.edu.au and hear our entire podcast archive, read articles and watch video content that explore the future of business.