Self-Driving Trucks Won’t Kill Millions of Jobs
In 2015, universal basic income advocate Scott Santens wrote a compelling piece arguing that self-driving transport trucks would kill up to 8.7 million jobs by automating the work of human drivers. The argument was picked up by a number of publications, helping to set off a narrative of fear around the future of the industry.
At the time, automation was in the news with frequent reports about the large percentage of jobs that would be lost in the coming decades due to the forward march of technology. In a very short period, millions of people would find themselves destitute. People like Santens had a ready answer: an unconditional $1,000 per month for everyone, which its supporters promised would solve a plethora of social ills.
As the looming threat of automation has faded from the headlines, it’s now clear that Santens’ argument played on the fears of ordinary people and the concerns of the moment to make a case for the policy that he makes his living promoting. We can now see that self-driving trucks are not the boogeyman they were presented to be. Not only is the technology further away than previously expected, but the number of people likely to be affected is much lower than Santens’ estimation — in the hundreds of thousands at the very most.
The warnings of basic income promoters and tech determinists are overblown.
Trucking jobs aren’t going far
In his argument, Santens seems to have used the highest numbers he could find to make his warning all the more threatening. He cites figures of 3.5 million professional truck drivers in the United States, with another 5.2 million people in industries supported by trucking — a very broad category that includes restaurants, motels, and “entire small town communities.”
The federal government, however, puts the number of heavy truck drivers closer to 1.8 million. It would still represent a major hit if they all lost their jobs, but that’s not what’s going to happen. Efforts at trucking automating do not target every possible type of trucking, but very niche aspects of the sector. Economists Maury Gittleman and Kristen Monaco recently published a rigorous assessment of the likely impacts of trucking automation, and their findings illustrate that the warnings of basic income promoters and tech determinists are overblown for several reasons.
First, not all heavy trucking can easily be automated. Long-haul routes with a lot of highway driving are in the sights of tech companies, but nearly half of drivers are employed for routes of 50 miles or less, meaning they’re not in much danger. In all, Gittleman and Monaco estimate that only 310,000 workers “drive tractor-trailers in the for-hire transportation and warehousing sector,” the most likely to be affected by automation. That’s a much smaller number than previously reported.
Second, it’s not just about distance and the type of route. Long-haul drivers perform other tasks besides driving, some of which aren’t as easy to automate. Among them are freight handling, paperwork, and customer service. Gittleman and Monaco also make the point that drivers act as security guards for the cargo. Freight theft came to $175 million in 2016, and it hasn’t yet been demonstrated how autonomous trucks would address that problem.
Finally, there’s a question of regulations. In the United States, state and federal regulatory bodies would need to allow driverless trucks to operate without human presence, and it’s certainly not clear whether they would allow that to happen. There are powerful interests on both sides of that legislative battle.
Gittleman and Monaco’s assessment throws cold water on the fears stoked by years of media stories on the likely outcomes of trucking automation. At most, 300,000 to 400,000 truckers would be affected, but likely not nearly that many. And the idea that 5.2 million jobs “supported” by trucking could also be lost is, quite simply, farcical. Businesses like gas stations and restaurants have a much larger customer base than long-haul truckers, and while they could be slightly negatively affected by a reduction in the number of human truckers, it wouldn’t be nearly as disastrous as has been presented.
But even if the worst-case scenario of 400,000 lost trucking jobs were to come to pass — and, again, it’s very unlikely — it won’t be for quite a while. Not only is the industry in desperate need of more truckers, but the tech industry’s timelines for autonomous driving were reset after a number of setbacks.
Autonomous vehicles vs. reality
Up until last year, companies could claim (with little pushback) that self-driving vehicles were only a few years away. There might be a few dissenting voices, but in general, the media would entertain their visions of ubiquitous autonomous shuttles and urban centers reoriented around them. But that all came crashing down in early 2018.
On the night of March 18, Elaine Herzberg was walking her bike across a stretch of poorly designed road in Tempe, Arizona. Little did she know, Uber was also testing its autonomous driving technology on a route that took it on that same stretch of road at that very moment. As the vehicle approached, it was unable to detect what she was and only determined 1.3 seconds before impact that it should stop. But it couldn’t; the team had disabled the car’s emergency braking capabilities, and they had failed to install a way for the car to notify the safety driver, who happened to be looking at her phone instead of at the road.
Herzberg died as a result of the crash, and the event served as a wake-up call for the industry. No longer could they believably claim that self-driving vehicles were only a few years away; instead, all the major automakers and tech companies began setting much longer timelines and reducing expectations for what the technology would be able to do. The key exception being Tesla’s Elon Musk, who many feel has lost credibility and is doing harm to the industry.
Automation isn’t a fait accompli nor are its outcomes.
A recent report by the New York Times featured quotes by executives on how they had “overestimated” the capabilities of autonomous vehicles. Argo AI’s Bryan Salesky said that while they’ve developed 80% of the necessary technology, the remaining 20%, “including developing software that can reliably anticipate what other drivers, pedestrians and cyclists are going to do, will be much more difficult.”
The future isn’t set in stone
It’s true that the technology needed to automate the highway driving of long-haul truckers may be a bit easier than vehicles navigating urban centers, but there are still a number of capabilities that are lacking and will require human drivers. For example, any heavy rain or snow is out of the question, and highways, where the road markings are faded, may need human assistance.
In short, truckers still have plenty of work, and those doing certain long-haul routes will likely be able to keep driving for decades to come. Trucking is a tough job, but a lack of work is the opposite of the problem it’s facing: The industry is actually short about 60,000 drivers, a number that could grow to 100,000 in the coming years as demand from Amazon and Walmart continues to grow.
Mass automation of trucking is a fiction developed by basic income advocates and tech determinists who wanted to scare people into accepting their uncritical prescriptions for the future. Automation isn’t a fait accompli nor are its outcomes. It’s being driven by corporate elites who care little for workers, and any positive future will require challenging their power, regardless of whether they’re ready to give everyone a poverty stipend.