The two Steves behind the Freakonomics franchise have just released another book, in which they apparently criticize self-driving vehicles. Unfortunately I currently have neither the time nor the money to read their book, so I’ll have to rely on this blog post from the Wall Street Journal, which seems to sum up their argument pretty well:
“Driverless cars could turn out to be a scourge on humanity.
They may lead to a worldwide surge in binge drinking since drunk driving would no longer be a worry. They also could be vulnerable to hacking by terrorists who send every self-driving vehicles in the western U.S. plunging into the Grand Canyon.
And by making car travel easier, driverless vehicles could lead to more congestion and pollution.”
They also cite the potential for self-driving cars to eliminate jobs (something I’ve addressed here) , though they also apparently admit that self-driving cars could reduce the number of car accidents.
I’m going to criticize the freaks’ approach to the issue, but I want to make clear at the outset that I don’t mean for a second to suggest that their predictions are implausible. They could be right. While the terrorism thing seems a little bit far-fetched, the result of this technology could be a world full of long-distance commuters in self-driving vehicles. This would contribute to vastly expanded urban sprawl, congestion, pollution, climate change, and oil resource exploitation. That is a plausible scenario, and it is a bad one.
Despite this, however, I think that the freaks’ logic is badly misguided. To illustrate why, it helps to point out that at the turn of the twentieth century, one of the most common destinations proposed that somebody would drive to with the newly invented automobile was the railway station. Cars, it was assumed, were mainly for local transportation or transportation through areas too remote to have their own rail lines. But if you wanted to go to a big city, then the train was your best bet. What this illustrates is that people always understand new technologies in terms of old ways of doing things. Historically, very few people have been able to accurately grasp the real implications of a radical new technology at such an early stage. That’s like trying to predict what the political landscape will look like in fifty years: It’s simply too complicated to accurately assess in advance. The best you can do is a lucky guess.
Levitt and Dubner, then, have probably made a mistake in assuming that the self-driving cars of the future will be used in same way that people use manually-driven cars today. This is not necessarily the case, however. Self-driving cars, in fact, open up a whole new range of possible behaviours. I’m going to explore one possible scenario, which I think is predominantly a good one, using two short speculative vignettes. The idea is to illustrate how everyday habits and lifestyle choices could be completely transformed by autonomous vehicles. Here’s the first one:
Julie was tempted to buy a car. That was, after all, what you were supposed to do when you got your first salaried job, and she did need to commute 20 miles every day to the office where she worked. But the autonomous taxi service that she had been using for the past year already served that purpose just fine. It meant that she didn’t have to drive or maintain a car. And the self-driving taxis were cheaper to boot. She decided she would rather spend the extra money on travel.
A huge part of the price of a taxi ride is the labour. That makes perfect sense, as taxi drivers need to make ends meet. But when you take the driver out of the picture, then suddenly taxis become a whole lot cheaper. This is especially the case if the self-driving taxis are electric, which will save considerably on fuel costs. I don’t pretend to know exactly what the price per mile of a self-driving taxi in the year 2040 will be, but it is plausible that it will be cheaper for the average person to rely on them than drive their own car. Self-driving taxis will be used far more efficiently than private cars could be; will be maintained by efficient, cost-optimizing businesses, and won’t have to pay for downtown parking.
This could start to make car ownership far less compulsory, even for those of us who live in the farthest-flung suburbs. Even if reliance on a self-driving cab is slightly more expensive than owning a car, it could still be a very attractive option due to the added convenience. And that means fewer cars on the roads.
“Alright, Andrew. We’re done predrinking. It’s time to go to the bar. I’m calling a cab”, said Stephen.
“Ugh. A cab? Do we have to? I’m kind of broke, and that’ll cost me like two beers”, replied Andrew. “It’s only three blocks away. We can make that in like half an hour”.
“Yeah, okay. Screw it”, said Stephen. “Let’s just walk.”
People who own cars tend to use them. Once you’ve paid the very substantial cost of a car, the marginal cost to use it to drive half a mile to the store becomes insignificant. So even if driving to the store is on balance a more expensive way of getting there than walking; this doesn’t translate into any direct perception of expense by the person deciding to drive to the store.
If, on the other hand, people rely on self-driving taxis, then the marginal cost of driving to the store suddenly goes up. Even with the vastly decreased price of a trip compared to a normal taxi, the fact still remains that the traveler has to pay upfront. And that means that many people will look to save money by walking or cycling for shorter trips. That means fewer cars on the road in total.
The ultimate outcome of these trends could be fewer cars on the street, more efficient use of the cars that remain on the street, and more electric cars. That would also mean more use of active transportation; more incentive to build high-density, walkable cities, and a more sustainable transportation system in general. As soon as you start to reduce the rate of car ownership, you reduce the incentives for car use. The neat thing about self-driving cars is that they allow a reduction in car ownership while allowing the same rate of car use to be maintained, at least temporarily. That makes them an ideal lever for changing the makeup of our transportation system and our cities, in addition to their recognized benefits in reducing drunk driving and accidents, and offering more travel options for the disabled.
Of course, my scenario is just as uncertain as the one proposed by the Freakonomics guys. Most likely, the way this will actually play out is completely different from either of our predictions. But my scenario suggests a few takeaway lessons. Firstly, it illustrates the value of a transitions perspective, which uniquely takes into account the complex ways that new technologies can interact with society. Secondly, it suggests a program for action. We can, to some extent, control the way that self-driving cars are introduced, through government policies and social advocacy. Self-driving cars could be good or bad in the end, but for the moment they’re an opportunity to change our transportation system. And we should think about how to take advantage of that.