Rating: High school and post-secondary
Summary: Markham interviews economist Jay Dixon of Statistics Canada about a series of studies “The Employment Consequences of Robots: Firm-level Evidence” and “The Effect of Robots on Firm Performance and Employment.”
This interview has been lightly edited.
Markham Hislop: Workers these days are worried about losing their jobs to robots and Statistics Canada has taken notice. They’ve done a study entitled “The Employment Consequences of Robots from Firm-Level Evidence.” To discuss that with us, we’re going to talk to Dr. Jay Dixon, an economist with StatsCan. In the energy sector, and particularly in oil and gas, we see automated ore haulers at Suncor’s oil sands mines that are going to replace 400 or 500 jobs. We see automated drilling rigs, which are moving into the industry now that will use fewer workers on the rig floor. Those are a couple of examples where we see robots are displacing workers, but maybe we could start this conversation with just a quick overview of your study, please.
Dr. Jay Dixon: The first challenge we had is actually deciding what robots are. There’s no real fixed definition. They are presumed to be machinery and equipment that have a degree of autonomy and a degree of decision-making ability that not noted in normal machines and are more the province of human beings. But there really isn’t a definite category that distinguishes machines with microchips or a degree of programming or a degree of intelligence.
We did notice that there was one category of data set that did seem to make that distinction: import data. Canada has a robotics industry, but it’s mainly focused on the programming and configuration of robots rather than the construction of the hardware. We were able to track robotic hardware coming into Canada and in many cases, we could attribute them to the firms that were using them.
Markham Hislop: Now, your study came to some very interesting conclusions about what type of workers get affected. I understand this correctly high-skill workers that have university degrees, their prospects improved, and low-skill workers with high school or less their prospects improved, but those in the middle those who maybe have a vocational degree or a trade they actually came out worse. Maybe you could explain that.
Dr. Jay Dixon: What we found is that when the robots came in, the firms themselves reallocated their workforces into different tasks. The high skilled workers go into design and customer service and other services surrounding the manufacturing process or the production process.
We weren’t really able to figure out what was happening in the low-skilled workers, except that there’s a lot of manual tasks that robots can’t really manage very well. And the general expansion of the firm encourages the hiring of those workers.
But the middle-skilled workers, the ones that had vocational training, often times I think has a routine aspect to their. work. One of the industries we noticed was chemicals. The lab technicians who were preparing samples and would have a vocational degree in order how to learn to do that precisely and properly, it was that type of task that required a degree of precision, but also had a high degree of routine, which is why the robots seem to be taking over.
Markham Hislop: Let me give you an example of the automated oil and gas drilling rig. I interviewed a couple of CEOs of companies in Calgary that are making these drilling rigs. And they said, that there are people on the floor, there are people handling pipe, manual labour type of work, those jobs are all going to be gone. Instead, there are going to be a couple of workers in the shack that’ll be supervising the computer.
That was a skill that was emphasized by those CEOs, a facility with data. They’re going to have to be able to supervise that machine and know what the numbers are saying and crunch the numbers themselves using software.
And the other thing was repair and maintenance because even though robots are more sophisticated, they require more maintenance and repair. And so there might be the same number of jobs, but they’ll be very different jobs. Is that applicable to some of the stuff that you’re seeing?
Dr. Jay Dixon: I think very much so, in addition, it’s one of the things that the robots have that machines don’t, is that they’re a lot more flexible, but you need somebody there to be able to teach them the tasks that they’re going to need to perform and then monitor them to make sure that the quality is consistent and help them when it’s not.
And those types of jobs definitely you’ll see a lot more workers tasked with becoming managers of the robots. The robots are doing the actual work and the people are monitoring and making sure that the robots are doing an appropriate job. With the middle-skilled people. a lot of their job already consists of using their judgment to make adjustments to the process that they’re doing. In this case, they are no longer hands-on, the robots are hands-on, but the workers are still there needing to manage and make sure that the robots are reacting properly to changing conditions.
Markham Hislop: That’s a really good point because the experts I’ve interviewed about this have pointed out that even if they shift jobs and they rely more on software and data, they still need that deep understanding of the industry because it’s about judgment.
Dr. Jay Dixon: Yes, and tacit knowledge is a very important part in a lot of these tasks that robots with machine learning techniques. Machine learning allows them to develop some tacit knowledge, but they’re still running considerably behind what the human brain is able to do. And picking up things that are noticing patterns that are not programmable, or are much more difficult to program.
Markham Hislop: One of the reasons why companies adopted robots is worker safety. That came up often in oil and gas and sometimes I have to admit I was a little skeptical, maybe a bit of spin by the companies to justify adopting the robot. But you found that in your survey.
Dr. Jay Dixon: Yes, though it varied across industries. For example, foundries put the robots between the human beings and the hot furnace, and to make sure that they weren’t being exposed to hazardous conditions. Another one was repetitive motion or handling awkward objects. Those types of skills were also a motivation in the construction industry – repeatedly lifting heavy objects – and with oil and gas exploration or nuclear power as well. You have robots that are going into hazardous environments, that human beings would be much more difficult to navigate safely.
Markham Hislop: Another thing your study found is that there will be fewer managers because once you’ve replaced the workers with robots and the workers are supervising the robots, then the workers themselves require less supervision.
Dr. Jay Dixon: Yeah. So we were very interested in this result and there’s two aspects to it. My coauthors and I have been debating what the primary motivation of this is, but one of the aspects is certainly because the robots in the course of their activity gather a whole bunch of data. You’re getting so much more information that the robot is gathering about the production process that allows you to make management decisions.
The other aspect is that we found that firms that adopted robots was to be more flexible. And so we’re wondering if maybe the flexibility means that you want more nimble workers, more empowered and also more capable of making quick decisions on the shop floor and less of a rigid hierarchy that takes longer to respond. We haven’t been able to disentangle those two issues, but those are two hypothesis that we’re thinking about it.
Markham Hislop: Just one final observation before I let you go, Jake. And that is it’s apparent to me from looking at the oil and gas industry and then other energy-related industries like wind and solar and hydro and utilities, that the very nature of work is changing. And the nature of work over the last, you know, from say 2010 is becoming quite different in 2020 and will be very different in 2030. Is that a fair observation?
Dr. Jay Dixon: As far as we can tell. Yes. I think that would be fair to say.