To meet climate goals, we can’t leave autonomous vehicles on autopilot
By Cameron Roberts, PhD
Autonomous vehicles have come to play a major role in many visions of the transportation future, with car companies, tech companies, and policymakers all proposing their own versions of a self-driving future. According to the most enthusiastic boosters, removing responsibility for driving from humans will not only make people’s lives easier and more convenient; but also make travel safer, more efficient, and less congested.
Even more potentially exciting, some predictions suggest that autonomous vehicles could radically reduce carbon emissions from the transportation sector. This prediction appears for several different levels of advancement for autonomous driving technology. Limited automation could increase the fuel efficiency of driving, while fully autonomous cars could be used to create a system of shared autonomous mobility on demand that could make car ownership itself obsolete, replacing it with a sustainable system of electric robo-taxis.
If these predictions come true, they would effectively be sustainable transportation policy on autopilot. Rather than grappling with the hard political, economic, and social dilemmas of low-carbon transportation, we could simply let a handful of clever engineers and entrepreneurs build a high-tech alternative that will gently supplant our existing system simply by being more convenient. Unfortunately, this tantalizing vision faces serious technical, economic, and political obstacles. In fact, there is also a very real risk that autonomous vehicles could increase carbon emissions from transportation. Both of these scenarios are addressed in a new report released by The Transition Accelerator.
Many of the potential efficiency gains from low levels of automation depend on how people will use autonomous vehicles in practice, rather than on the technology itself. “Eco-driving”, for example, could take advantage of precise digital control to make driving patterns more efficient—particularly if a majority of cars on the road are autonomous and if these cars are networked with each other. Car owners, however, might prefer to not use such a feature due to concerns about privacy and personal autonomy that are well-established in the research on this subject. This literature also reveals a devoted group of “hard-core” motorists, for whom the physical act of driving is extremely important.
The potential benefits of automation get larger at higher levels of automation, but so do both the challenges and the risks. If autonomous vehicles can drive themselves entirely independent of human intervention, then drivers could redesign them from the ground-up, for example by emphasizing efficiency over performance and by taking advantage of safer driving to remove bulky safety features. At the same time, however, it could lead to a huge increase in the total distance travelled by car. A two or three-hour commute might look less unappealing if people can work, watch TV, or even sleep en route. People might take advantage of the ability to multitask while driving to travel much longer distances, which would mean more cars on the road, more congestion, and more emissions. They might also start preferring larger and less efficient vehicles where they can stretch out and fully take advantage of their new free time while underway.
The most radical scenario for autonomous mobility is one in which fully autonomous vehicles mean that car ownership is replaced by shared autonomous taxis. Companies such as Waze and Uber are already banking on this possibility. There is, however, considerable evidence from surveys that most people would not be willing to give up their cars in favour of such a system, and would prefer to use privately-owned autonomous vehicles.
This could be partly because a car is not just a means of travel. It can also be a mobile living room and storage locker, and a way for its owner to display their class status. Self-driving taxis might therefore mainly compete with public transit and active travel rather than private cars, thereby resulting in higher carbon emissions
A major hazard associated with fully autonomous vehicles is the potential for empty vehicle-kilometers. People could send their car home rather than parking it downtown, or even dispatching it to run errands independently. This would result in a colossal increase in the total vehicle-kilometers travelled by car, which would in turn increase carbon emissions.
Many of the scenarios sketched out here are pessimistic ones, but they are plausible enough to be cause for concern. Even if the negative impacts of these scenarios are offset by other innovations, such as electric vehicles, they will still make the challenge of responding to climate change harder rather than easier.
None of this means that we should reject autonomous vehicles out of hand. It is still possible that autonomous mobility could play a major role in a future low-carbon mobility system, for example as a last-mile solution for public transit, or in the form of small, slow, shared micro-shuttles like those being piloted in Europe under the City2Mobil program. This, however, will require hard work. To make a real contribution to mitigating climate change, it is not enough just for autonomous vehicles to disrupt the mobilty system; they must disrupt it in a way that acts as a force-multiplier for other forms of sustainable mobility, such as electric vehicles and public transit.
This means that autonomous vehicles will not create a low-carbon mobility system for us on autopilot. It may be an ironic conclusion, but to have the greatest benefits for the climate, autonomous vehicels will have to be manually steered.