Who is driving? Computer is driving.

There are conflicting visions of a future with autonomous driving. One version seems to focus on a select few benefits: expanded availability to disabled, elderly, and perhaps even children, vastly reduced traffic deaths (they were on the order of 30,000 in 2014), and the possibility that you won’t have to own a car at all: you’ll simply summon one when you want one. An article from The Atlantic’s City Lab seems to focus on this latter point and its potential impact in disrupting public transit systems: why would anyone ride public transit at all?

Yet as I think about a possible future with driverless cars, it seems more likely that people will continue to own their vehicles, as this is the primary way the median American travels to and from work. There are probably deep cultural reasons somewhere in there too. The article has some interesting notes, and it seems to me that very few people are thinking about how autonomous driving technologies would implement into public transit systems in cities.

In fact, if driverless cars encourage a large class of people to ride in cars more frequently (they no longer have the pain of driving themselves, and can focus on work or some other form of leisure in the car), then things like commute times and congestion (and possibly pollution, absent greater gains in emissions technology) may become an even greater problem. The only solution to this problem, electric cars or not, autonomous driving or not, is increased investment in public transit.

So what happens to public transit? Potentially, implementing this autonomous driving technology could eliminate or greatly reduce the share of labor costs in public transit. A bus can be added to a route without hiring a person with a pension, health benefits, and overtime restrictions. This could be good from the perspective of a commuter: costs would be more predictable, service could be expanded, and chronically underfunded systems could potentially have more cash to invest in infrastructure and maintenance. Adding a bus route used to depend on the fixed cost of potentially buying a bus and the variable cost of adding a worker to drive the bus for certain hours of a day. Without a worker, the bus would sit idle. Now, the sole essential cost is a fixed one (or the role of labor costs would be much smaller, perhaps limited to inspection and repair). Furthermore, pre-set bus routes seem more predictable and easier to implement in the autonomous driving algorithms as compared to the many additional challenges faced with a personal car, where the destination could literally be anywhere. Fewer edge cases would be encountered.

This is one possibility. Yet another is that companies like Uber, which are investing heavily in a driverless future (few people mention the dark motivations of this: Uber doesn’t want to deal with the headache of having actual employees), could take advantage of the current state of affairs in public transit. It’s not hard to imagine that the acrimonious relationship between many cities and Uber could be changed: tired of dealing with labor unions in the public transit sector and chronic under-service, they make a deal with Uber to start providing some bus-like services. And so the privatization of government services in American society may continue. This may seem crazy right now, but I have to think that companies like Uber and Lyft are thinking beyond their current model. They have already thought of car pooling, and they are already investing in driverless cars to rid themselves of the headache of dealing with actual employees. It may not be crazy to imagine Uber getting into operating its own line of buses as cities turn desperate.

Another interesting question that comes out of all of this: what happens to the jobs? Public transit systems are heavily unionized, and the relationship with the public is ambivalent in many ways. A job that provides generous benefits and a living wage should be applauded: on the other hand, in a society where median wages have stagnated and living in many urban centers has grown more expensive, fewer and fewer think “we should try and get what they have”, and instead think “if we don’t have it, neither should they.” Having lived through several public transit strikes where service was largely shut down by unions, bringing city life to a standstill, I have some sense of the complex feelings involved.

So then in 30 years, when the occupation of driver (whether bus, taxi or truck) has disappeared, where do these millions of workers transition to? It has been a fallacy throughout history to decry technology as destroying jobs on net: new industries arise from the gains in technology and productivity and fill the gap created by the destroyed industry. But so far, I have seen very few imagine what this future would look like, and where these jobs might appear. Instead we hear about a future where the Silicon Valley unicorns and monopolies gradually solve all of society’s ills. Furthermore, even if new industries appear, we have learned over the past few decades just how hard transitioning careers can be, and how permanently harmful the process can be to one’s lifetime earnings (not to mention psychological well-being). Research by David Autor has also shown that jobs have become increasingly bifurcated: a hollowing out of middle-class jobs and an expansion in low-wage jobs, and to a lesser extent, highly skilled jobs. And has the role of technology fundamentally changed, moving from something augmented by labor to something completely substituting for labor?

If it’s going to happen, it’s going to happen. But the future to aim for is one where institutions are robust and able to handle the future’s potentially massive job losses, and one that doesn’t envision a personal car for every person who can afford it, exacerbating congestion issues and disadvantaging the poorest. As the population grows and technology becomes potentially more disruptive to the workforce, investments in infrastructure and accessible education should be made a priority. Out of all the claims made by Silicon valley types about the huge technological advances just around the corner, autonomous driving seems to be the one that might actually come to pass. It’s better to be proactive if we know the changes are coming.