Autonomous Vehicles and Technological Unemployment in the Transportation Sector
29 January 2014
Technnological unemployment has been back in the news in a big way. There was a widely reported study, The Future of Employment: How Susceptible are Jobs to Computerisation? by Carl Benedikt Frey and Michael A. Osborne (cf. Half of jobs to be lost to computerisation?), and recently The Economist devoted a cover to the topic (Coming to an office near you…), with several stories inside the magazine considering technological unemployment from a variety of perspectives.
I have visited the question of technological unemployment several times, most particularly in the following posts:
Developments that touch upon technological unemployment — the actual automation technologies, our understanding of these technologies, and the conceptual infrastructure employed in attempting to make sense of economic and technological trends — evolve so rapidly that it is necessary to revisit the question on a regular basis, even while sedulously remaining focused on the big picture so that we do not mistake some passing and ephemeral trend for a development that will define a new era of history.
The big picture of technological unemployment is that it is part of the ongoing industrial revolution, which changed the relationship of human beings to each other and to the planet they occupy, and which continues to unfold with unprecedented developments. Some who write about the industrial revolution make a series of distinctions between the first, second, and nth industrial revolutions, but none of these finer distinctions have been universally recognized, so they only tend to create confusion. Indeed, when I was reading an article about technological unemployment last week the writer called technological employment the second industrial revolution, either unaware or unconcerned that others have already called previous developments (as, for example, electrification or assembly-line production) the second industrial revolution.
I prefer to think of the industrial revolution as one, long, unfinished process, beginning in England with the use of fossil fueled steam engines to power machinery and changing continuously up to the present day, as new technologies emerge from the previous generation of technologies. This technological innovation that began the industrial revolution and sustains it in our time I have called the STEM cycle. Because of the STEM cycle, the industrial revolution continues to revolutionize itself, always producing new technologies and new technological dislocations in the socioeconomic system, but it is the structure of technological change itself that defines the industrial revolution and the industrialized societies that have arisen in its wake.
The same conditions that held in the earliest automation of formerly manual tasks continue to hold today: some tasks are easier to automate than are other tasks, and some parts of a given task are easier to automate than other parts of the same task. The automated production process tends to break down tasks into their simplest constituents, automate the automatable tasks, and then stitch together the whole in an assembly-line production process in which the gaps between automated tasks are filled by human workers who continue to perform the tasks (and parts of tasks) that cannot be readily automated. Thus industrialization gives us the “job of the gaps” employment structure, and continuing technological innovations narrow these gaps, reducing employment.
However, even as entire classes of employment disappear, new classes of employment appear — as unprecedented as the technological innovations that spelled the end of previous forms of employment — and this has allowed industrialized economies to continue their balancing act of keeping the majority of their populations employed while enjoying the rising productivity that results from continuous technological improvement of the production process. However, there is no guarantee that this balancing act can be maintained indefinitely — or even that this would be a desirable state of affairs. Imagining a permanent future of dead-end industrial jobs is a kind of dystopian scenario that offers little hope. However, the utopian scenario of human beings freed from stultifying labor by technological unemployment seems too good to be true.
I will discuss some of the implications for technological unemployment in relation to the transportation industry, since I know something about the transportation industry, having earned by income in the industry for the past three decades. The rapidity of the development of self-driving cars (autonomous vehicles) is a testament to the rapid gains of technology and computerization as they bear upon transportation. When, in the past, people imagined an automated road transportation network (and this is a staple of futurist thought that has been imagined many times), it was assumed that radio transponders would have to be built into roads and infrastructure to guide a vehicle along. Instead, laser range finders and radars construct a local map of the terrain, which is then compared to high resolution maps of the actual environment, and the precision of GPS systems allows the vehicle to navigate through the map. (Of course, it’s a bit more complicated than that, but that’s the abbreviated version.)
The development of autonomous vehicles is a potential boon to the transportation industry. One of the greatest challenges to the industry has been the ability of motor carriers to find a sufficient number of drivers to haul their loads, and recent hours of service (HOS) regulation changes have increased the limitations on the number of hours a driver can drive in a day and in a week. Autonomous commercial vehicles, when they become both practicable and legal, would potentially mean unlimited freight capacity and trucks operating twenty-four hours a day, seven days a week. Driver shortage would no longer be a problem for freight haulers.
While most driving is routine and could easily be handled by an autonomous vehicle, there is a significant portion of the driving day which is likely to elude automation for some time to come. Driving a tractor-trailer within a congested urban area is much more difficult that driving a passenger vehicle in the same conditions, and it will take longer to automate this process than the hours on the open highway between major urban centers. There is nothing in principle that cannot be automated, and when the technology is available it is likely that autonomous tractor-trailers will be safer in traffic than human drivers. However, a single severe collision involving injury or a fatality would likely be picked up by the media and one would expect a headline something like, “Killer Robot Trucks on our Highways!” This would likely to delay the development of the industry for years, if not decades. Due to the obvious liability issues, one would expect that the technology would not be rolled out until it is fully mature, and even then accidents will happen. (I have elsewhere argued that industrial accidents are intrinsic to and ineradicable features of industrial-technological civilization, and traffic accidents are among the most common of industrial accidents.)
Other than the complexities of driving in crowded urban conditions that put other drivers, cyclists, and pedestrians at risk of life and limb, there are aspects of freight hauling that will not be easily automated. Another aspect of our industrial-technological civilization is that it runs clock around and year round. There is never a break. Freight moves every day of the year, and if the transportation infrastructure is slowed or stopped, store shelves are quickly bare. Other than unpredictable snow storms that shut down highways, there are predictable inclement weather conditions that occur on all roads at high elevations. In the continental US, thousands of trucks every day go through mountain passes, and it is not usual in the Rockies or the Cascade Range for drivers in mountainous areas to chain up their vehicles every day simply to be able to complete their trip. Tire chains are a nearly archaic technology, but they are effective, and nothing else will get a truck through snow and ice like chains. Believe me, I’ve been there. I know whereof I speak.
I think it will be a very long time before any robot or automated system will be able to chain up a tractor-trailer in inclement weather conditions. There are automatic chains available, but their use is limited, and they won’t get you through deep snow. Putting tire chains on a tractor-trailer is physically demanding and difficult to do well. No doubt there is a way to automate the process, but it won’t happen in a robust form any time soon — and here by “robust form” I mean a dependable way of getting a truck through a mountain pass on a daily basis.
I can foresee a day when tractor-trailers are automated for long stretches of highway in flat country, and dual-purpose vehicles are sometimes piloted autonomously and sometimes driven by human drivers. It might be possible to station drivers on the outskirts of cities, who would then get into autonomous vehicles and drive them within urban areas. Or drivers might be stationed at the bottom of mountain rangers, and get in the trucks to take them over the pass. But in a severe winter, the snows come down the side of the mountains, and the stationing of drivers to take over in inclement conditions might have to change daily. Under such conditions, it would be an open question as to whether it would be more cost effective to simply keep drivers in the trucks all day rather than attempt to constantly shuttle drivers to where they would need to be to take over autonomous vehicles where these vehicles could no longer safely operate. So truck drivers aren’t yet quite out of a job, even when autonomous tractor-trailers become a reality.
The process of automating commercial vehicles is likely to spread out over many decades, which will allow for realignment of employment within the industry over time. And driving, of course, is not the only job within the transportation industry. There is the warehousing and loading of freight, maintenance of vehicles, and many other functions. It will be a very long time before automated roadside service for breakdowns will be possible. Autonomous vehicles will be more technologically complex even than the trucks on the road today, and they will break down with some regularity (breakdowns, like industrial accidents, are an intrinsic part of industrial-technological civilization). Automated vehicles broken down on the shoulder of the road will have to be serviced by human technicians for many decades to come, and a stranded automated vehicle would also be a soft target for cargo theft, which creates a new kind of “opportunity” for human beings within an automated economy.
. . . . .
…and so it begins… Dutch Trucks Will Drive Themselves …note in relation to the above that there are no mountain ranges in Holland.
. . . . .
. . . . .
. . . . .