Speaker: Cameron Roberts

Date: February 22nd, 2021

Cameron Roberts is a postdoctoral research fellow at Carleton University. Cameron works with The Transition Accelerator on electric mobility and mobility on-demand, as well as on other transportation technologies that can radically disrupt the car-dependent transportation system in Canada. His background is in the history of transportation, having studied the history of aviation, railways, and automobility during graduate studies at the University of Edinburgh and the University of Manchester. Cameron has been a recipient of the Darwin Trust of Edinburgh PhD scholarship, and has published articles on the history of transportation and sustainability transitions in Energy Research and Social Science, Science as Culture, Technological Forecasting and Social Change, and Technology Analysis and Strategic Management.


There is a lot of hype surrounding autonomous vehicles and the changes that they could bring. One term that is starting to show up in a lot of writing on this topic is the idea of a ‘Holy Grail’ consisting of autonomous vehicles as part of a shared mobility-as-a-service system. Further exploring the potential of this new autonomous vehicle system is important because the implications are potentially revolutionary. The critical outcome that Cameron investigated in this presentation is what this technology means in terms of its environmental impacts, specifically greenhouse gas (GHG) emissions and climate change.

To provide background, Cameron opened up with a discussion of what autonomous vehicles are, their history, and contemporary rhetoric surrounding them. Though predictions vary, some believe that, in the future, people may outlaw driving cars, because of the radical safety improvements that come as a result of autonomous vehicles. In fact, up to 95% of the automobile miles travelled in the US could be on-demand autonomous vehicles.

The general logic of this hype surrounding autonomous vehicle transport is as follows. First, there is a recognition that the current system of car ownership is extremely inefficient (financially, spatially, temporally). This is because most cars are not used most of the time, and even when they are used, they are rarely at full capacity. Therefore, shared autonomous vehicles would be a cheaper and more efficient option than driving privately-owned vehicles. This trend towards shared autonomous vehicles could spell the end of private car ownership and the rise of “robo-uber” or “robo-taxi” systems due their financial advantage. Furthermore, due to economies of scale, these autonomous vehicles would be overwhelmingly electric and there will be fewer of them (as they are used more efficiently). Finally, the result of these changes would be huge improvements in safety, traffic efficiency, time use, land availability, and GHG emissions.

Some researchers have argued that this process by which our transportation system could be transformed is self-driving and already underway. In fact, there are indications that this process is already underway within car companies and tech firms. If we allow the process to reach its conclusion, it could solve our modern-day mobility issues. This is key because changing the mobility system has proven to extremely challenging; it is very politically, culturally, and socially difficult to get people to change their mobility habits.

Despite optimism, Cameron argued that there is reason to be wary. There have been visions of autonomous vehicles for almost as long as there have been cars. Experiments and demonstration projects of autonomous vehicles can be seen as far back as the 1920s. Though this initial research was rudimentary and did not lead to notable change, autonomous vehicles would return in earnest at the turn of the 21st century. By this point, electronics and computing ability had drastically improved, and thus much more progress could be made. A big turning point for autonomous vehicles was November 2007, when the American Defence Research Agency (DARPA) did the Grand Challenge III, which challenged autonomous vehicle innovators to create an autonomous vehicle able to navigate an urban environment. Several contestants succeeded, and this was an important proof of concept step for autonomous vehicles. Just two years later, in 2009, Google began working on autonomous vehicles. Shortly after, the first law permitting the operation of autonomous vehicles on a state’s road was passed (in Nevada, 2011). From the 2010s to now, both car companies and tech companies have jumped on the autonomous vehicle bandwagon, making significant progress in the field.

Though the hype around autonomous vehicles has fluctuated since the 2007 DARPA Challenge, there are several ongoing persistent technical challenges that the technology still faces. These challenges include poorly-marked roads, unpredictable pedestrian behaviour, inclement weather, hard-to-recognize road obstructions, and a lack of reliability of the hardware and software. These challenges must be addressed for there to be widespread adoption of autonomous vehicles.

Cameron then went on to discuss how autonomous vehicles fit into discussions of climate change, specifically the challenge of reducing emissions associated with our transportation system. This climate context is key to the discussion of autonomous vehicles, because it sets the demand for what we need to accomplish with this technology. Globally, transportation systems are responsible for 23% of GHG emissions. The vast majority of these emissions are from road transportation, meaning that addressing road transportation emissions is a key aspect of addressing climate change.

Beyond climate change, another important factor driving this transformation of our transportation system is air pollution, which is responsible for nearly 15,000 premature deaths in Canada each year. While not all air pollution is from transportation, it is a key contributor to this problem.

Despite the potential of a fleet of electric autonomous vehicles to address these climate and air pollution issues, predictions regarding the outcome of such a transport system vary. Some predictions suggest that this would radically reduce emissions from transportation, but others indicate that autonomous mobility could lead to a massive increase in vehicle-kilometres travelled, and an overall increase in emissions. Thus far, most of the academic literature considering these possible outcomes has not seriously engaged the social science research studying which outcome might be more likely to occur in practice. Therefore, Cameron addressed this gap, bringing together the technical, energy and climate, and social science literature to answer the question of what this autonomous vehicle technology means for carbon emissions in the transportation sector.

Despite ongoing challenges, electric vehicles are being increasingly adopted. But several barriers remain if this new generation of vehicles are to be an effective climate change mitigation strategy:

  • Electric vehicles are not inherently carbon-neutral
  • Fossil fuel vehicles sold today will remain on the road for an average of 10 years
  • Manufacturing electric vehicles produces a lot of carbon (embodied emissions)
  • Even the most optimistic forecasts of electric vehicle uptake fall short of would be necessary to mitigate the worst of climate change

The key takeaway from this is that if autonomous vehicles cause people to start driving a lot, electric vehicles are not going to address the climate crisis.

However, the effects of autonomous vehicle uptake on our society go beyond emissions, many of which are difficult to predict. For example, maybe increased vehicle safety from autonomous vehicles will cause people to drive faster, which could lead to increased air pollution. Because of these less predictable and often unintended consequences, predictions regarding the impact of autonomous vehicles on GHG emissions vary greatly. Some predictions estimate that emissions could be reduced by almost 100%, while others predict that emissions could increase more than 200% due to an increase in vehicle-kilometres travelled.

To provide insight into which of these predictions is correct, Cameron reviewed three scenarios: limited automation (e.g., driver assist features), full automation (completely autonomous driving), and full autonomy (cars that drive themselves with no human present). In the first scenario, autonomous vehicle technology is limited to technologies like parking assistance and collision avoidance. Widescale adoption of this limited vehicle automation could lead to several benefits, such as reducing traffic congestion, which is a major factor contributing to carbon emissions from transportation. Limited vehicle automation can also lead to increased comfort for passengers since accelerations and decelerations would be smoother.

However, a lot of motorists do not find the idea of autonomous vehicles attractive at all as they enjoy the experience of  driving .There are also significant privacy concerns with autonomous vehicles. To get the maximum benefit from this type of transportation system, the vehicles would need to communicate with each other and with the road infrastructure. This data could be accessed and used for various benign or nefarious purposes. Lastly, there are economic inequality issues with autonomous vehicles. For example, there are already proposals to use market-based software to auction off places at autonomously-controlled intersections. In such a system, you could bid money to reserve at place in this intersection so you could drive straight through without having to stop. In situations where there is a lot of competition at this intersection, wealthier people will be able to move much faster than poorer people. There are also concerns surrounding the rebound effect and induced demand (autonomous vehicles could reduce congestion, which could lead to increased driving).

So, what does this all mean? Cameron suggested that limited automation could lead to marginal improvements in emissions, or a marginal worsening of emissions depending on a variety of factors. Full automation could lead to substantial benefits, such as a massive, coordinated traffic intelligence system which could radically reduce congestion and increase road safety. But this could also allow drivers to multitask while travelling, which leads to issues of induced demand (reduced perceived cost of driving). Thus, while there are substantial potential gains associated with this fully automated transportation system, there are also potential downsides. The final scenario Cameron presented was that of full autonomy, the so-called ‘Holy Grail’. In this scenario, people may stop owning vehicles altogether, since autonomous electric taxis would be cheaper and more convenient. Particularly if these taxis are shared, this could result in vast efficiency improvements, cost savings, and emissions reductions. The potential benefits are huge in this fully autonomous scenario, but there are several issues. These issues include vehicle cleaning costs, personal preferences for wanting to own a private vehicle, and security features for passengers. This system of shared autonomous mobility can have negative impacts on the climate as well: if travel is cheaper, people may travel more (induced demand).

Because of all of these issues, more optimistic scenarios focus on a targeted use of autonomous vehicles. For example, using smaller autonomous vehicles as a ‘last mile’ solution to support public transportation. These vehicles could also serve to support other low-carbon forms of mobility, such as cycling and walking, through providing transport for the occasional long distance or heavy load trip.

Cameron concluded his presentation with a note of ambiguity. The narrative of using autonomous vehicles to create an efficient, coordinated, and hassle-free transportation system is in no way new. While technical improvements may mean that this dream actually becomes reality in the near future, there are reasons to be skeptical of this narrative. The most likely benefits of autonomous vehicles, particularly in the short term, are marginal and will not be enough to meet climate goals. There is much more uncertainty surrounding a more drastic transformation and its potential associated benefits, and this would require heavy and controversial policy interventions. Therefore, autonomous vehicles are only a Holy Grail if we work hard to overcome the risks associated with them, and the interventions required to do so would be extremely politically contentious. In short, autonomous vehicles are not a simple shortcut around contentious mobility politics.


Q: In the case of completely autonomous private vehicles, are there negatives in cyclic one-way traffic (i.e., to and from work) where the vehicle would return home after dropping you off so it would double the mileage?

A: Yes, absolutely. That is a huge part of the risk of that scenario. The literature calls it ‘empty vehicle kilometres’, and that is one of the most common fears. It is a side effect of one of possible benefits of autonomous vehicles, that you do not need parking downtown (which takes up a lot of space). You end up with empty vehicle kilometers in a shared autonomous vehicle model as well since these must travel empty to pick up and drop off locations.

Q: How far along are we (with regards to approving the use of autonomous vehicles) in the transport/freight sector in Canada? Is this sector treated differently from a policy standpoint than the lightweight car market? Is there current opposition to this transformation  from truckers?

A: I have not investigated that in too much detail. I think in terms of the timelines for approval my suspicion is that it would not be too different, or possibly trucking would take a bit longer than passenger cars because a truck is a more dangerous vehicle. I doubt we will get widespread use of these before 2030.

Q: Are  fully autonomous vehicles only feasible in small cities such as Singapore? Would implementation in larger cities that are more car-based like Los Angeles would be more difficult?

A: There has been research on this, and autonomous vehicles could actually do a pretty good job in rural areas, particularly if we’re discussing autonomous vehicles as a supporting force for public transit. A single big transit line could go through a few rural communities and then an autonomous vehicle fleet could help people travel that last mile to get home.

Q: You’ve made a compelling case for why autonomous vehicles could be detrimental for climate, safety, and congestion – you also argued that there’s huge hurdles in the way of fully automated vehicles taking over our roads. However, when powerful players can possibly gain from a sociotechnical transition, they can use their access to resources and decision-makers to make it happen. I am curious if you think that there are players who stand to gain from a high uptake of private autonomous vehicles, and if it is possible that those players would push them through all of the regulatory, cultural, and technical hurdles for their own gain?

A: That is an interesting question, and I am working on a paper on this right now. Transportation has always been political and there have been massive political upheavals over changes of transportations systems at multiple points in history. We currently have a large coalition that supports the basic paradigm of our private car-based transportation system, including the fueling system, the licensing regime, and vehicle sellers. But the tech industry is throwing everything they have at this, both in the technical sense and in a political sense. So this is going to be a struggle, with powerful players on both sides.

Q: How does the electric vehicle ecosystem adapt in the event of an extreme weather event which affects all the aspects of mobility and communication? would it leave the system very susceptible?

A: I have not thought about this in detail, but I can think of a few responses. First, the telecommunications infrastructure will be damaged, which would cause problems. Second, extreme weather could lead to unpredictable obstructions in the road, such as tree trunks. There could also be issues with the electricity supply, for example what is happening in Texas right now. In this scenario, you would not be able to charge the electric autonomous vehicles.

Precis completed by Silke Popescu, MA Sustainable Energy student.

Event Page 

Seminar Poster