Man vs the machine: How automation can nudge the employment balance of USA's teetering trucking industry

In this regard, Parth Vaishnav, an assistant professor in the Centre for Sustainability at the University of Michigan and Anirudh Mohan, a graduate research assistant in the department of engineering and public policy at the Carnegie Mellon University, have co-authored an academic paper entitled – Impact of automation on long haul trucking operator-hours in the United States.

The researchers relied on statistics from the Commodity Flow Survey (2017) dataset publicly available from the US Census Bureau, jointly released by the US Bureau of Transportation Statistics, and the US Department of Commerce to examine the impact that automation would have on hauliers' operator hours, particularly along highways. The paper propounds that automated trucks (ATs) will impact "up to 94% of operator-hours for truck drivers….if the technology is deployed across the continental US in all conditions."

According to the researchers, this in turn would lead to job losses, and a disturbed political climate rife with protests. The paper attempts to clarify the 'conflicting claims' surrounding ATs: "The promise of ATs has led to widespread concern about job losses in long haul trucking, which is a common profession in the US, particularly for men with high-school education (Wertheim, 2020). On the other hand, it has also been noted, often by the companies developing this technology, that long haul trucking currently faces a labour shortage and automation will create new short haul jobs, which will more than make up for the long haul jobs lost. As a result of these conflicting claims, as well as the uncertainty over the technology itself and its limitations, there is little clarity on how automated trucking will be deployed and its economic and political ramifications, such as the impact on the long haul trucking labour market."

For the operator hours affected, the duo propose a 'transfer hub model' which may lead to alternative job sources for the displaced truckers. The transfer hub model mentioned in the paper draws on an idea originally proposed by Jeffrey Hickman in the National Science Foundation (NSF) Workshop Report: Effect of Automated Trucks on the Truck Driver Workforce co-authored by Hickman, Frank Levy, Stephen Burks, Steve Viscelli and John Lee in 2018. According to the research paper, the transfer hub model "would involve an automated truck (AT) completing the highway leg of the route and human drivers undertaking the more complex suburban-urban segments at both the starting and end points of the journey. Truck ports near highways would be used to switch out the trailer from the prime mover and enable this switch at both ends."

However, the proposed transfer hub model may potentially not be enough to fulfil the jobs lost in the trucking industry. Moreover, the researchers also encountered a reluctance from the truckers' side, "to shift to new modes of operation such as short haul driving that will result from any application of the transfer-hub mode of deployment."

Image Source: Impact of automation on long haul trucking operator-hours in the United States, Feb 2022. Originally published in Nature Journal

The researchers further cite from the 2018 NSF report, quoting Steve Viscelli who estimates nearly 300,000 long haul jobs to be at risk due to automation. "As per the analysis, these at-risk jobs are primarily concentrated in dry van and refrigerated trucking, which are characterised by high turnover and low wages," the paper states.

To understand the dynamics and the findings better, The STAT Trade Times reached out to Parth Vaishnav to unravel the social consequences of the impacts of long haul trucking.

What spurred the study?
Anirudh and I work in automation and we were particularly interested in the energy use of cars. In the course of our studies, we came to realise that the energy to automate the truck is particularly significant on the highway, so it is fairly mature. We wanted to further delve into the energy impact aspect of this. It's actually possible to automate this, and that is how this paper originated. We wanted to understand how much of it can be automated. That is the question that we started out trying to answer and we're still trying to answer.

You have also devised a transfer hub model that has to be operationally less complex when it comes to the highway. The study also considers automated trucking tests which are currently only limited to the sun-belt region of the USA. In this context, could you take me through some of the challenges particularly when it comes to the transfer hub model?
So, the transfer hub model includes automation, but even without automation, there are certain benefits. Basically, the trucks right now are a compromise between what is optimal for driving on the highway and what is optimal for driving in urban settings. They're currently not particularly good at doing either of those two things. For a transit hub model, you can design a truck that is really optimal for city and for highway driving and that might give you environmental benefits and cost saving as advantages of the transfer hub model.

In terms of automation, currently there are limitations as to where the technology is being deployed, which means that the technology will get better over time. Through our paper, we try to break that out as to what happens if you deploy the truck only in the sun-belt states or what happens if you only deploy it in good weather, or what happens if you only use it for really long trips? We ran the analysis for all of these aspects. And that is the benefit of actually kind of breaking down the work at the individual scale. So, if you look at minutes of job losses from automated trucking, treating the entire truckload industry as one component and then estimating what proportion of the jobs are actually lost. We scaled it down, and took a closer look at individual trips or journeys and tracked where those occur.

There was a proposition in your paper about how if the transfer of deployment is done in sites which are different from the truckers' place of stay, that could pose a severe problem as around 40% are from rural areas. Could you elucidate on this?
So one of the advantages of being a long haul trucker is place of stay doesn't matter. You can live in a place that is low cost, alternatively, you can live near the mountains or somewhere because the only time you're at home is when you're not working and when you're working, you're on the road. In the US, it is not unusual for long haul truckers to be away for a week as they don't just go to some place and drive back. They go from one place to another, so they might do a big loop around the country, and that keeps them away from home. There are obviously costs to being away from home so much, but there are also benefits because your home can be wherever you want it to be and it's not tied to your work. One of the major benefits of the transfer hub model is that you now probably can be near cities. So for a lot of truck drivers, that would be an attractive proposition. There'll basically be a shift to the cities, because of the transfer hub model.

The paper makes mention of another proposition on how partial automation is viewed negatively by heavy-truck operators. Is it mainly because of safety concerns?
If you have a system which is automated for the most part but where humans have to monitor it and quick interventions are necessary if something goes awry, the resulting situation is very stressful for any human. You are basically entrusting your life to a computer. For long distances, basically one trucker drives and the other truck rests and drive in turns so you don't have to stop and can cover really long distances.Even in those kinds of situations, truckers say that they find it really hard to sleep when someone else is driving because if that person messes up then they could be injured or killed. So, entrusting your life even to another human is difficult for truckers and adding technology into this equation is something truckers find really difficult to accept.

Considering the policy aspect, there was mention of how policymakers should ask truck operators to basically reinvest and ameliorate the disruption to employment as well as reduce environmental footprint. So if you can just take me through a little about the environmental aspect involved.
In the state of California if you want to operate an automated vehicle on the public roads, you have to make sure that the automated vehicle is zero emission. This is where states and governments have a bargaining chip on their permissions to operate on public roads and there are enormous benefits. There's an enormous financial benefit to operators to automate and regulators can say that well, if you want to drive this as yet risky technology on the public roads, then the public should benefit. So I think that is one point of leverage. This is going to save trucking companies a fair bit of money, and it's quite reasonable that some of that money be reinvested to help truckers.

The paper mentions that an increase in short haul operations is unlikely to compensate for the loss in long haul operations. Is there any other way that this can be compensated in terms of the truck ports?
It's hard to say that it's impossible, but there would be a fair amount of work involved. The proportion of short haul driving is the percentage of total hours. So, there's just no way that the 5-10% of additional hours that you create could make up for the work that might be there at the truck ports. It is not clear that employee would have the same skills or would be paid in the same way as trucking. For instance, if a person is allotted to participate in automation tests, that person does not have the same skills as a driver, who in turn, does not have the same skills as that person. The other thing that one needs to be a bit careful about, which we've mentioned in the paper, is that automation involves a loss of work. The reason we said a loss of work rather than a loss of jobs is that at least in the US, when it comes to long haul trucking, truckers often are on the job for less than a year. So, when you say that the job is at stake, you have to ask, what is actually lost? This is not a segment where someone has been in training for many years. So, where the trucker may have had many years of stable employment in the job and in automation there is no employment is not an ideal scenario.

The paper mentions that the threat of jobs lost is likely to trigger political disturbances. What kind of disturbances might these translate to?
It is the case that normally when there are job losses, people protest. We can expect that this may be the case for trucking as well. We don't necessarily know what forms those protests might take. There is literature which suggests that when you introduce changes in technology that result in job losses, people are not happy and may register protest.