Advanced autonomous vehicle technology is maturing at such a rate that most observers expect it to be ready for commercial deployment in the very near future. Some companies have already begun live trials of semi- and fully-autonomous vehicle platforms on public roads. The general safety and productivity advantages of these developments is undeniable, but little discussion in any public forum has occurred regarding the negative human and social impacts of widespread deployment of commercial autonomous vehicles (CAVs).1
Specifically, commercial trucking, widely cited as the likely first “target” for CAVs2, directly employs 3.5 million drivers, and 4 million more via indirect support industries.3 The successful deployment of CAVs will squeeze these workers’ livelihoods by making drivers far less essential to commercial transport, and then crush them completely once the technology reaches its “final” form. This is not happenstance, or theoretical; it is the primary objective of their deployment.
What I will motivate and discuss here is not whether autonomous vehicles should be developed or deployed; their advantages are too great and too obvious. However, given the well-known history of how the unfettered automation of entire industries has previously devastated entire regions and generations, I believe we must take steps to make the transition to autonomous vehicles one that benefits all of us without irredeemably harming some of us.
Not “just a job”
Driving a truck isn’t just a job…it’s a career with its own culture and communities that supports millions of families across the country. Pay grades nationwide average $40,000 – $50,000 yearly4; this is triple the national hourly minimum wage of $7.25/hour5, and nearly double the poverty level for a family of four6.
This makes truck drivers far more like other well-paid workers whose livelihood has been devastated by automation in the past (auto factory workers come to mind) than less skilled, less well-paid, and more transient roles (e.g. cashiers, which have been getting replaced by self-checkout facilities for years). Further, an unemployed truck driver will have few opportunities to find work that pays similarly: absent other skills, they will be forced into looking for work paying far less than their former profession.
Aside from individuals’ monetary concerns, truck drivers are an essential part of the United States’ character. Driving a truck is the most common job in the country, and as Planet Money‘s visualization graphically illustrates, it is hardly a localized profession; driving a truck is the most common job in 30 states as of 2015:
Of course, these factors together make truck drivers an obvious target for “improving productivity”.
The Machines are Coming?
This is an overused catchphrase used in news reporting about and editorializing on the latest wave of automation eliminating jobs — often well-paying jobs, these days. While the pieces themselves are usually more nuanced, the catchphrase implies that the automatons approach of their own volition, or perhaps exist and advance as an unyielding force of nature.
No, the machines come because we collectively decide that they should. Sometimes by explicit intentional choice, but often because we passively accept the choices made by others, usually corporate-government entanglements organized to maximize productivity and profit.
For all the reasons why driving a truck is a stable and rewarding career as noted earlier, it is also a plum target for elimination by more efficient, more productive, less costly technology. According to the American Transportation Research Institute, the research arm of the largest trucking industry advocacy group, one third of the cost of every mile driven is attributable to driver wages and benefits.7 Though CAVs will cost more than traditional rigs, eliminating drivers (after first squeezing more miles out of each one thanks to initial, less sophisticated CAVs eliminating many of the human constraints around hours of service) will drive costs down and maximize miles driven and thus profit made per day.
We don’t need to infer this motivation, though: none other than the CEO of Uber (which also owns Otto, a leading developer of CAV technologies that is partnered with Budweiser, conducting live tests of their Level 4 retrofit8) has publicly declared that their ultimate intention is to replace human drivers.
What will the impact of this be? I’m no economist (yay!), but for perspective, the number of potentially-affected commercial truck drivers (remember, 3.5m) is approximately the same as the number of jobs lost in 2008 in the U.S. as a result of the financial services collapse and resulting “Great Recession”, and over 40% of the number of jobs lost throughout the entire recession (8.5m).9 Of course, the deployment of CAVs will take much longer than the duration of the “Great Recession”, but this is cold comfort to the drivers and millions of others that will be affected.
“They should do something”
Depending on your political bent, you might now be expecting or hoping that those in positions of power were thinking long and hard about how the negative impacts of CAVs could be mitigated.
Unfortunately, existing regulation and government oversight of autonomous vehicles is either entirely absent, or singularly focused on safety and liability concerns. I am not aware of, and could not find any references to any existing or proposed regulation to address the significant dislocation in the commercial driver labor market that would surely occur if Level 48 autonomous vehicles were to enter widespread use.10
Peeking at the underbelly of the policy world makes clear that the labor impact of CAVs is thoroughly ignored. A meticulous 2016 200-page report on autonomous vehicles written for policymakers by the RAND Corporation contains a single lonely paragraph to acknowledge the catastrophic impact CAVs will have on drivers’ livelihoods11:
Jobs will also be lost. The act of driving is the source of many reasonably well-paid jobs. Recent immigrants often operate taxicabs or livery services, and municipal bus operations are the source of many union jobs. The commercial transportation sector employs thousands of professional drivers. Just as the invention of the automatic elevator led to the loss of many operator jobs, it is likely that AV technology will eventually lead to the loss of commercial transportation sector jobs at considerable human cost. Ultimately, the lost jobs might be replaced by others, perhaps related to the AV industry, but there may be considerable economic disruption.
It’s particularly painful to see a gross inaccuracy regarding the number of jobs at issue (“thousands” cited in the RAND report, when the number of commercial truck drivers actually tops 3.5 million) coinciding with blind hope and faith in creative destruction yielding economic benefits for others.
While public policymakers sleep on the worker impacts of CAVs, autonomous vehicle vendors are actively arbitraging what little regulation and oversight does exist in order to secure the lowest possible degrees of governmental “interference” in their development and live testing. Uber is again leading the way here, with their recent deployment of autonomous vehicles in San Francisco despite known safety problems, which were rapidly relocated to Arizona when the state of California ordered the vehicles halt testing on public streets while Arizona’s governor opened theirs without restriction. Far from atypical, these tactics are actively encouraged, even by investors in Uber’s competition; for example, Mr. Srinivasan is a Board Partner at Andreessen Horowitz, an investor in Uber competitor Lyft:
The universal sole priority demonstrated by CAV vendors, their investors, the trucking industry, their customers, public agencies, and policymakers is to accelerate and maximize the deployment of CAVs, with the resulting massive profits accruing solely to CAV vendors and owners. Meanwhile, nothing is being done to prepare or protect truck drivers from the disruption being planned for them.
A modest proposal
There are probably dozens of policy proposals that could improve the outlook for truck drivers and those that depend on them. I have one in mind.
The dislocation of millions of truck drivers losing their livelihood will come bearing costs; my fundamental proposal is that these costs should be borne by those that benefit from the deployment of CAVs, and the greatly-improved profit they promise.
There are many personal, familial, and social consequences of job loss that cannot be parcelled and countered by any compensation; but, some are. In order to avoid the widespread social unraveling widely documented as a result of automating other industries, under- and then unemployed truck drivers will need:
- unemployment benefits and other transitional assistance
- access and admission to suitable education and job training programs
- health care and life insurance to ensure their families’ well-being
The objective must be to get those affected by CAVs transitioned into other quality work. Not doing this will mean that millions of truck drivers will likely fall into poverty along with their families, impacting their general welfare for generations. Without public policy addressing the unique threat that CAVs pose to truck drivers’ livelihoods, these needs will either be paid for out of already strained general unemployment funds and low-income assistance programs, or not provided for at all…with tragic consequences for everyone involved.
State motor vehicle registries should create a new registration class for CAVs (Level 4 and above8), with a yearly registration fee commensurate with the estimated costs associated with the benefits necessary to maximize the chances of a dismissed driver landing safely.
Of course, a blanket unchanging registration fee would represent an unreasonable barrier to the eventual (and again, very desirable) widespread deployment of CAVs. This can be addressed by adjusting the special CAV registration fee over time; it could:
- be low for the first small percentage of a state’s cumulative registered trucking fleet, so as to encourage the early adoption, testing, and refinement of CAV technologies
- treble up sharply for the large middle bulk of a state’s fleet. This will allow the state to build a sizable fund from which to pay benefits to displaced truckers, and should lengthen the horizon of mass deployment, hopefully allowing most truckers and others in supporting industries enough “warning” to exit to better opportunities while they’re still employed.
- drop to zero and end special treatment of CAVs for the last portion of a state’s trucking fleet, thus opening the floodgates for the full benefits of CAVs, and maximal profit for their manufacturers and owners thenceforth.
Autonomous vehicles have tremendous promise, and will one day make our roads far safer and more efficient. However, a significant portion of that efficiency is predicated on the elimination of millions of decent-paying jobs that will help a far fewer number build and grow fortunes. We should celebrate the innovation that enables those fortunes, but we must not let those that suffer in its wake to do so alone and unaided. It is incumbent upon us, as members of a just society with history and memory of prior adventures in yielding the fruits of automating once well-paying work solely to the automatons’ creators and owners, to not repeat those prior errors of inaction.
- The same can be also said of non-commercial autonomous vehicles: their arrival is universally described as ushering in a utopian future of safer, more efficient, less unpleasant personal transportation. As far as I can tell, there is little to argue with there, compared to the large and specific threat that CAVs pose to a large and essential population of current workers. Thus, the focus here on CAVs and their impact. ↩
- While Uber has publicly declared its intentions to eliminate all of its drivers by introducing autonomous vehicles, the economic impact of doing so — while still painful for those impacted — is far smaller. According to the Bureau of Labor Statistics, there are just 234,000 taxicab drivers and chauffeurs are employed in the U.S., each of which earn on average $11.00 / hour. These figures are dwarfed by the economic footprint of commercial truck drivers, which makes them a far more appealing target for improved efficiency measures, so to speak. ↩
- American Trucking Associations, ‘Professional Truck Drivers and the Trucking Industry’, June 2016, http://www.trucking.org/ata%20docs/what%20we%20do/image%20and%20outreach%20programs/misc%20documents/pro%20Truck%20Drivers_final.pdf ↩
- Bureau of Labor Statistics, ‘Occupational Employment and Wages, May 2015: 53-3032 Heavy and Tractor-Trailer Truck Drivers’, https://www.bls.gov/oes/current/oes533032.htm ↩
- U.S. Department of Labor, minimum wage topic summary, https://www.dol.gov/general/topic/wages/minimumwage ↩
- U.S. Department of Health and Human Services, ‘2015 Poverty Guidelines’, https://aspe.hhs.gov/2015-poverty-guidelines ↩
- American Transportation Research Institute, ‘An Analysis of the Operational Costs of Trucking: A 2014 Update’, http://www.atri-online.org/wp-content/uploads/2014/09/ATRI-Operational-Costs-of-Trucking-2014-FINAL.pdf ↩
- The different levels of autonomous vehicle capabilities are defined by SAE International Standard J2016, summarized in http://www.partnered with Budweiser, conducting live tests of their Level 4 retrofitsae.org/misc/pdfs/automated_driving.pdf. Level 4 autonomous vehicles (those that are able to negotiate most driving tasks in most conditions without any assistance or attention paid by a human in the cab) will be the first that will significantly affect truck driver employment. Lower levels of vehicle autonomy are thoroughly beneficial to all stakeholders, as they are focused strictly on safety, and require an alert operator at all times. ↩ ↩ ↩
- The BLS maintains this data (see the Excel/CSV download link at the bottom of the page if you want to tinker with it), or you can refer to https://en.wikipedia.org/wiki/Job_losses_caused_by_the_Great_Recession for easier reading. ↩
- The threat to the commercial truck driver has been cited in arguments for universal basic income, including https://medium.com/basic-income/self-driving-trucks-are-going-to-hit-us-like-a-human-driven-truck-b8507d9c5961 ↩
- Anderson et al., ‘ Autonomous Vehicle Technology A Guide for Policymakers’, published by the RAND Corporation, 2016, http://www.rand.org/pubs/research_reports/RR443-2.html, page 39. ↩