I’ve seen many articles about self-driving cars getting the go-ahead to expand in cities/regions like San Francisco and Austin. These announcements have been met with mostly positive reactions from the public because of the obvious benefits of self-driving cars. Self-driving cars are safer than human drivers, and prevent stress created from driving. My worry, however, is that these benefits will not be as widespread as advertised. While the safety issues are difficult to argue, adoption will be very slow, and the general public will still have similar costs associated with transportation. In addition, it can be argued that massive adoption of self-driving vehicles will introduce new safety converns. It’s possible that self driving cars will only have a safety benfit (if that) with little to no social benefits or monetary benefits.
As a thought experiment, let’s suppose that self driving cars are legal everywhere tomorrow and self driving taxi companies like Waymo have the infrastructure to accomodate 50% of all the active cars on the road. Tesla self driving will not be included because there are still questions about its safety. Here are my thoughts on what will happen day 0, 3 months in, 6 months in, a year in, 5 years in, and 10 years in.
Day 0. Not much will change, some beta testers of self-driving cars are happy to use self-driving cars, but it will take a while before there is still a lot of adoption.
3 months. About 20% of people who commute to work in a car are using self-driving cars. The rest of the commuters are still hesitant of the technology or find it cheaper to drive themselves. There is still rush-hour traffic since not enough self-driving cars are on the road. The number of drunk or fatigue induced accidents is lower, by about 25%
6 months. More adoption for daily commuters, maybe 35%. For certain workers, it is still cheaper to drive their own cars, or they work in remote locations where self-driving vehicles require 20 minutes to pick them up. There is still rush-hour traffic. The number of accidents related to fatigue and drunk driving is 50% lower than before self-driving. Self driving cars are disabled on days with low-visibility & rain. Their behaviour is too unpredictable and has caused unnecessary accidents during rush-hour traffic. Most self-driving cars are still only electric, so the public has range anxiety about using self-driving cars for road trips. In addition, there have been more cases of self-driving cars that cannot handle extreme random road situations like poorly marked detours due to construction or making way for emergency vehicles in tight road situations.
1 year: Finally we hit 50% adoption for daily commuters. There is still traffic during rush hour. Self-driving cars still cannot handle rainy or low-visibility days. There are more reported instances of self-driving cars unable to handle 1% of driving situations.
5 years: Self-driving companies have scaled up their infrastructure so 75% of cars on the road during rush hour are self driving. There is still a ban on self-driving cars during rainy days and low-visibility days, although the number of days where the ban takes effect is less than 20 days a year. With 25% of cars on the road during rush hour still being self-driving cars, there is still traffic, so your commute is only 10% faster. The number of different destinations of self-driving cars within a downtown make self-driving cars are suboptimial. Articles and research is getting published that to really improve traffic with self-driving, all cars on the road need self-driving.
10 year: The benefits of self-driving have not improved over the last 5 years. Some people still own some type of vehicle that they can drive themselves to feel some kind of freedom. The auto industry has promised to only make cars with self-driving capabilities because of their popularity. The number of fatigue induced or drunk driving related accidents is effectively 0. There are still cars on the road, causing noise pollution, and people still need a car to do many basic tasks like get milk.
Many kids in suburban areas go to school without their parents in a self-driving car. They see only a car seat in front of them, or they are looking at some new Iphone 25. They have no sense of their environment or what roads they need to take to get to school. All new infrastructure that is built has identifiable markers for self driving car cameras to make self-driving safer.
I hope in sharing this thought experiment, people can realize that self-driving cars will not fix that much. Rush hour traffic will still be a problem, and the general public, while initially pleased with self-driving cars, will still feel the effects of isolation, loneliness, and cost that comes form requiring a car to go anywhere.
Self-driving cars will make driving safer, and I will not argue that point here. What I will argue, however, – at the risk of sounding like someone who think abstinence means “safe sex” – is that if less people are driving, then overall, driving is safer. Yes, this is a blog post about urban planning. Self-driving cars will not get cars off the road, so rush hour will still exist. People will still need to open an app on their phone to go anywhere. Eventually, the general public will get tired of this and want independence from self-driving cars back, and buy a car they can drive themselves again.
I love self-driving cars, but I don’t want to loose sight of a bigger goal which is better urbanization. We need denser cities with more public trainsportation options. I lived in San Francisco, a pretty urban city, and at times, I still felt isolated because only 3 busses served my neighborhood, and they were infrequent and unreliable.
Self driving will not solve traffic, nor will it solve the innefeciency of every family owning multiple cars.
I hope to release a blog post about how a denser city & better public transit & better bike infrastructure can help solve these problems in a much cleaner way.