The supply and demand whiplashes of the Covid-19 pandemic snarled global supply chains, shaking up labor markets and well-established migration patterns in the process. Existing cracks in logistics and infrastructure systems widened, thereby making them newly visible. In the US and Europe, a dramatic shortage in the supply of long haul truck drivers sparked panic among businesses and policymakers in 2021.
In her new book Data Driven: Truckers, Technology, and the New Workplace Surveillance, Karen Levy of Cornell University offers an in-depth view of the US long haul trucking industry, explaining why so few workers today are willing to take up what was once considered a respectable, skilled job. Decimated by waves of deregulation and union-busting since the 1970s, a once highly organized and well-paid workforce has fragmented over time, subjected to the intensifying discipline of markets and management.
Over the past decade, a new technology has emerged to threaten the autonomy of truck drivers: the electronic logging device (ELD). ELDs, which became mandatory in 2017, replace the fallible, driver-completed paper logs which have long allowed drivers and companies to subvert the “hours of service” regulations which place limits on working time. But Levy compellingly shows that ELDs function simultaneously as regulatory devices, instruments of managerial control, and objects of worker resistance. Overall, they appear to undermine the skills and autonomy of truckers to the point that they singularly fail in their stated aim—reducing driver tiredness and accidents—while at the same time opening up new ways for truckers to undermine the authority of managers and regulators.
Levy further scrutinizes some of the automation technologies coming down the line. Puncturing hyperboles about fully autonomous vehicles peddled by Silicon Valley’s prophets, she argues that “human/machine hybridization,” rather than a jobs apocalypse, is the most likely scenario over the medium term. Due to the impossibility of eliminating human labor from the complex bundle of rote and safety-critical tasks performed by truck drivers, she foresees an intensifying rollout of bundled automation and surveillance technologies. The “cyborg” trucker of the near future will face brainwave, eye pattern, heart rate, emotional, and other kinds of biophysical monitoring by a range of wearable and in-cab devices. Ostensibly to ensure safety, such devices serve another purpose—cementing managerial power over performance management and control.
An interview with Karen Levy
WEI WEI: Your book discusses the impact of technological integration on workers in the trucking sector. What led you to this topic?
Karen levy: I am a lawyer and a sociologist by training. I’ve always been obsessed with rules, and how rules are used in the world. Recently, I’ve grown interested in the use of technology to enforce rules—either by making it more difficult for people to break them, or by surveilling people and monitoring who breaks them. We have moved to digital rule enforcement in all kinds of domains, either because we think it is more cost effective, or because we think it is more consistent. But when we do so, we also discover that there were good reasons that rules were imperfectly applied to begin with—that is, that rules as they live in society are not as simple as they appear.
In graduate school, I spent some time searching for an arena in which to study this transition from manual to digital rule enforcement. By coincidence, I heard a radio broadcast about the digital enforcement of “hours of service rules” in the US trucking industry, which determine how long a trucker must drive before taking a break. That day, I decided to visit a truck stop and speak with some truckers there, whom I found incredibly thoughtful and interesting. That led to my dissertation, and, ultimately, this book.
steven rolf: The US trucking sector is notable on the one hand for its prevalent libertarian ideology, and on the other hand for its legacy of militant collectivist organizing through the Teamsters. How do you square these apparently contradictory impulses in the workforce?
Kl: On the history of unionization in the industry, the key person to read is Michael Belzer. His book Sweatshops on Wheels is indispensable to everything I think about trucking and collective action. It is true that trucking was heavily unionized in the mid-twentieth century—about 60 percent of the industry was unionized in the early 1970s. But trucking was also one of the chief industries hit by the deregulatory wave of the early 1980s. During this time, the federal government abolished barriers to entry and stopped setting standard freight rates. This prompted a race to the bottom in an increasingly competitive industry where companies were trying to cut costs in order to offer discount rates. Among the key things that were cut were the conditions of labor; the annual salary of a truck driver tanked from $110,000 annually in 1980 to $47,000 today. Union membership rates also declined dramatically; they were down to 25 percent in the 1990s and have continued to fall since. Today, it’s hard to find a unionized long haul truck driver.
I think this deregulation is responsible for the contradiction you outlined—truck drivers work long hours under dangerous conditions and are undercompensated for their work. This is both what decimated truckers’ unions and contributed to this libertarian ideology.
WW: Digital surveillance of workers is increasing across various sectors. For example, in warehousing there is digital real-time monitoring through the use of wearables, and in banking, digital surveillance is used to prevent insider trading and ensure regulatory compliance. Is there something distinct about the digital surveillance of trucking work? Or can we see some sort of convergence in how digital technologies are integrated into broader systems of social control?
KL: I ask myself all the time whether there is anything special about the sort of surveillance we see in the trucking industry. Is it just one instance of some broader dynamic? In some ways, what truckers are experiencing resembles what workers in warehouses, food services, and professionalized industries like law and medicine have seen. You could even say that trucking is just catching up with the monitoring in other low wage sectors. Because trucking is mobile and geographically distributed, truckers have been able to maintain autonomy and preserve immunity from the surveillance common in factories, offices, or warehouses.
But there are some things that distinguish trucking and the digital surveillance which takes place in the sector. One is that trucking constitutes a very unique workplace—truckers live in their cabs for days or even weeks on end, which is very different from entering a workplace and leaving to go home. If you talk to truckers about why they do the work that they do, they will often tell you that they didn’t want someone looking over their shoulder all the time. In this context of total independence, surveillance strikes at the core of a deep occupational identity. It is very deeply connected to their sense of self and their self worth.
Another key issue is what I call surveillance interoperability. In trucking, some of the data collection is mandated by the federal government through those timekeeping regulations I mentioned earlier. But that government monitoring, which is actually not so extensive on its own, serves as a scaffold for corporate surveillance from companies. Now that these companies had to buy and install these electronic logging devices, they may as well use them to fulfill their own organizational goals. These include fine-grained performance monitoring of drivers—how much fuel they use, how hard they break, how fast they drive. There are cameras trained on their faces to see if their eyelids are fluttering and so on. This is easy and cost effective for companies because it also operates for government data collection. On top of that, there is third party data collection—this data is very interesting to third parties who want to do things like sell parking spots to drivers. In the book, I talk about how these different forms of surveillance are mutually constitutive and interoperable not just technically, but economically, culturally, and legally.
WW: In the book, you describe what trucking might look like in the future. Can you describe how the relationship between automation and surveillance is evolving in the trucking sector?
KL: In the 1960s, Manfred Clynes and Nathan Kline wrote about the cyborg and the idea of technology as an augmentation of the human, giving humans greater control over their environment. In the workplace, that integration between humans and machines has had the opposite effect, where technology has been used to more closely supervise and control workers.
For the future of the industry, I think we can continue to expect that automation and surveillance will have a complementary relationship. There is a lot of fear over worker displacement—via autonomous vehicles for instance—but we are not nearing this reality for a number of technical, social, legal, and economic reasons. In the short term, the role of artificial intelligence and automation is to hybridize the human trucker forcibly with machines, with wearables, cameras, and other technologies. This relationship between automation and surveillance is something we can expect to see in all kinds of contexts. Even “autonomous” systems require the human to interact with the machine in some way, and this will result in surveillance over that worker.
SR: We’ve recently done some research on the trucking industry in Brazil and China, in which we found that services like route planning and automated pricing are far more widespread in middle income countries than advanced ones. Just yesterday, I read an interview with the CEO of Uber Freight in which he said that they account for 2 percent of all freight moved in the US market. Why have these digital disruptors made such quicker strides in poorer economies?
KL: It definitely seems to be the case that services like Uber Freight seem to have more of an influence in middle income countries. That might in part be due to how concentrated the long haul trucking market is in the United States. Brazil and China have much greater concentration of ownership in small fleets, and those types of carriers can really benefit from things like load matching—matching unassigned loads to carriers with available capacity. In the US, 80 percent of assets are carried by 20 percent of trucking firms, which is to say that there are some big firms that are really dominant.
Load matching could be useful for owner-operators—where drivers lease or own their trucks—and small fleets, but many of the drivers in these arrangements find themselves in exploitative situations, like long term lease-to-own agreements. This means they can often only drive for a particular company, and they’re beholden to that company for deciding what they’ll carry and when. Steve Viscelli’s The Big Rig explains how companies keep workers in really precarious positions where they have neither the employment protections that an employment arrangement might afford them, nor the freedom to decide what or for whom they’ll haul. I think these constraints on autonomy are part of the reason digital companies have been less successful in the US.
SR: From our perspective, it is interesting that platformization has so far predominated in non-critical industries. You can platformize take out food or a cab. But for critical logistics, we are reaching an entirely new level of co-ordination problems. Assets and people need to be in the right place at the right time, so the degree of coordination is far more demanding. How far do you think this can go? Can the entire industry be reliably organized on a just-in-time basis?
kl: That is very difficult to imagine given the current political economy of the industry. The coordination costs may not be so high, but the incumbents in the industry do have a lot of power. When we think of the number of assets and people that need to be utilized, it feels like a very distant future.
SR: At the beginning of the book, you observe that despite the truck worker shortage in North America and Europe, firms continue to provide low quality jobs which results in high employee turnover and labor market dropouts. How will automation impact the value of trucking labor?
kl: More experienced workers tend to object more to greater digital supervision, and it’s not difficult to understand why. These workers have enjoyed a great amount of autonomy for years and have millions of miles of safe driving under their belt, when suddenly they are told they have to start doing things completely differently because they are no longer trusted to perform their job correctly. Many truckers I spoke to said that in-cab monitors made them feel like criminals or children.
The younger truckers have fewer objections. They’ve never done it any other way, and there aren’t many jobs requiring relatively little training that are not subject to deep workplace surveillance. So, ironically, the more experienced workers are driven out by these technologies which are meant to promote safety.
When it comes to automation, people like to suggest that workers who lose their jobs can be upskilledinto leadership positions. But anyone who thinks this is going to happen in trucking hasn’t spoken to many truck drivers. People’s occupational identities matter, and we can’t pretend that they don’t. When we think about automation and reskilling, we need to remember that people aren’t cogs, they carry histories and occupational pride. You can’t just reconstruct the identities people build over long periods of time.
SR: Your book does point to some examples of resistance. But overwhelmingly, it seems, these avenues for resistance are being closed off with greater turnover and the declining occupational identity you describe.
kl: I think you’re probably right that it has become more difficult over time, but there is still a fair amount of informal informational exchange among truckers—at truck stops, on message boards—where drivers can exchange some of these strategies.
When writing the book, I found it surprising the degree to which resistance is sometimes accepted or even encouraged by companies. Companies want monitoring and compliance with federal rules, but they also want stuff to move quickly. If that requires breaking the law, they will encourage truckers to break the rules. To some degree then, we can expect resistance to persist in part because it serves corporate interests.
SR: I’ve been struck by the reactionary nature of recent collective action in the North American trucking sector, from the 2019 Black Smoke Matters anti-regulatory protests that you describe in the book, to the more recent Canadian trucking protests against vaccine mandates. Do you see any prospect for a shift towards more progressive mobilization?
kl: We have seen some progressive action when it comes to unionization rates among workers for delivery tracking companies, like UPS. There, people work locally and get to know one another. Things are more complicated in the long haul segment, where the work is geographically distributed and fairly isolated. The culture is consequently very focused on preserving autonomy; in my own conversations with truckers I didn’t find a single one who seemed excited about collective action.
In order to think about progressive actions for workers in this industry, we have to look to other mechanisms. One of the most impactful tools in this regard has to be a reform of the pay structure to make sure truckers are paid for the work they do. At the moment, truckers are exempt from the Fair Labor Standards Act, which provides access to things like workplace protections and overtime pay. This would allow us to address things we tend to sidestep like monitoring fatigue—you remove the incentive to overwork if you pay people well for the work that they do.
Ww: One of your key arguments in the book is that technology on its own is not deterministic. Factors like culture, economy, and institutions matter. What role do these other factors have in protecting workers, promoting the public interest, and upholding human dignity?
kl: That is the million dollar question. I think the most important thing we can do is recognize that when technology seems to be used to solve a problem, it is often actually being used to avoid confronting an even deeper problem. This is sometimes called the digital band-aid or digital duct tape—technology keeps things together enough while not addressing the root causes of a social, economic, legal, or cultural issue.
Truckers are a good example. They are subject to all this fatigue monitoring because they’re overworked, overtired, and underpaid. But instead of resolving these underlying elements, we manage the situation using digital surveillance. I’ve also done some work in collaboration with a gerontologist about monitoring in nursing homes, where people put monitoring devices in the rooms where their elderly relatives live. This technology is perceived to be necessary because people don’t trust nursing home facilities, and that is because they are underfunded and understaffed.
The tricky thing about technology policy is that it’s not really about technology at all. The center of required policymaking is often in the economy or through social organizations, and technology might have an important role in how we address the problem. But we should use technology as kind of a lens on that broader social landscape rather than as a substitute for resolving underlying issues.