Self-Driving Cars Aren’t As Environmentally Friendly As You Think

Tesla, Google, Apple, and Uber may be building self-driving cars, but a new study shows their ecological benefits are overhyped.

The endgame of Uber is a utopia where you step out your door, and one pod in an entire network of self-driven vehicles picks you up to ferry you to your destination. There’s no stop and go traffic! No jamming on the gas like Vin Diesel! No car ownership! Surely, this self-driving culture will be better than our planet, right?


Not necessarily.

In fact, depending on a complex series of factors, the environmental costs of self-driving cars could neutralize their benefits, according to new research published today in Transportation Research Part A by University of Leeds, University of Washington, and Oak Ridge National Laboratory. Their paper analyzes a multitude of theories about emerging autonomous vehicle technology–like whether a safer car could be lighter, and thereby more fuel-efficient–and scales them to the real driving habits of cars and trucks across the U.S.

“A couple years ago, there started to be a lot of enthusiasm and hype around this technology,” says Don MacKenzie, Assistant Professor at University of Washington. “Our sense was there were a whole lot of people throwing around benefits–like cars that can drive in [aerodynamic] formation–but we thought, how much can that actually help you? Let’s put some boundaries on it.”

The result is a no-nonsense view of the benefits that we may or may not get from self-driving cars–and it includes some serious surprises.

The biggest shocker is that a world without traffic jams wouldn’t move the needle on energy efficiency all that much–only lowering energy costs by 0-4%. “I would say the most overhyped benefit from energy standpoint is congestion mitigation,” MacKenzie says. “Relieving congestion can do great things for people’s time, but it’s not doing much for energy or emissions.” The reason is that, while it may feel like we’re all stuck in traffic all the time, the reality scaled across U.S. roads is that we’re really not, and fuel costs from traffic congestion is relatively low. Traffic jams account for a mere 2% of all fuel cost today. And so while stop-and-go traffic can reduce a car’s fuel economy by as much as 50%, a United States without traffic jams would “save maybe 5%” on the total energy tab of driving by 2050, MacKenzie says.

The team is more bullish about other potential benefits of autonomous vehicles, though. Take the idea of Uber matching each user with the appropriate car in terms of size. Their analysis found that if car manufacturers supported the idea with single-seat cars, and no one owned their own private vehicle, this strategy could cut back on fuel usage by nearly 50%.


Another big potential gain could be found in “platooning,” which you may know better as drafting, in which large flocks of vehicles could be coordinated to drive on highways together, reducing aerodynamic drag across the group. Platooning could lead to as much as a 25% reduction in our vehicular energy use. However, it’s also a prime example of how complicated these projections are to generate, the authors explain. The efficiency of platooning depends on how close self-driving cars can cruise in relationship to one another, how long these trains of vehicles could grow, and whether urban infrastructures could adapt to take advantage of platooning in denser areas.

No one knows the real answers to these questions yet, and only the realistic application of these technologies coupled with public policy will truly answer them. Platooning, in the worst case scenario calculated by the paper, might offer only a 4% reduction in energy usage. That’s not any more helpful than the aforementioned gains offered by traffic mitigation!

“With further work, we can start to narrow these ranges,” MacKenzie says. “But what we were trying to do here was point to which are the biggest levers, and which have the most uncertainty.”

All of this said, the saddest takeaway of this research is that, despite the benefits that self-driving cars may bring our environment, those gains will inevitably be offset by the fact that we’ll be more likely to take a car somewhere when we no longer have to drive. If we aren’t bound by our steering wheels, and we can socialize or do work in our cars, as companies like Ideo have envisioned, public transit becomes a lot less enticing.

“It turns out that a driver’s time is the single biggest cost of driving a vehicle,” MacKenzie says, noting that it outweighs a car’s purchase price, fuel costs, and even accident risk. “The intuition is basic economics. You make something easier and more convenient, and people use and consume more of it. You make vehicle travel easier, you can expect them to travel more.”

And therein lies the fatal flaw with the perfect, self-driving car: We’d actually want to use it.


About the author

Mark Wilson is a senior writer at Fast Company. He started, a simple way to give back every day.