Awhile ago, I wrote a “controversial” post about the impending robotopia. I thought I’d follow up on that with a prime example of the incredible efficiency gains that could be seen from automating an industry, plus the inherent downsides and the inevitable fight to stop it from happening. I hope it will become clear that the gains far outweigh the costs, and that we should figure out another way to mitigate the effects of the costs so that we can reap the huge benefits of this developing technology. I’m of course talking about the self-driving car (do try to contain your surprise, I know that took you off guard).
I think that most people view the self-driving car as a gimmicky luxury good or concept car, similar to the electric car before Tesla came along (sidenote: the Tesla model S is a magical space car and I think everyone should own one), but the truth is that this concept showcases the mind-bending transformative power of automation. There are of course obvious benefits — imagine reading the paper as you drive to work! Woo, how exciting. But apply that concept to your bar-hopping trip last weekend — no matter how drunk you are, the computer driving the car isn’t. And all of a sudden the self-driving car has saved 10,000 lives a year in this country alone. And it keeps going; almost every one of the 30,000+ traffic-related deaths in this country each year is due to human error.
We’ve already transformed the world, and we’ve only seen the most obvious improvements; we haven’t even scratched the surface of the potential of this technology. The inefficiencies of our current system stem from the fact that each car is driven by a single human with imperfect knowledge. Now imagine a fully-autonomous, networked transportation system. Cars would be able to know the future routes of all nearby cars and act accordingly; completely eliminating the potential for nearby drivers to do something unexpected means you can pack cars closer together and they can travel faster. Knowing what’s coming up ahead means cars won’t succumb to every commuter’s nightmare, the traffic wave. Moreover, knowledge of global congestion patterns and application of a locational marginal pricing scheme (think HOT lanes, but everywhere) that allows computers to take the least-cost route to a destination (or even propose alternative destinations that avoid congestion), and you’ve significantly reduced congestion on the roads. But wait! There’s more! You can suddenly change the rules of the road if the entire fleet of domestic transportation is automated. We all agree to travel on the right side of the road, for instance, and so heavily-trafficked roads have many lanes in each direction, and often only half of them are in use as traffic flows predominantly into or out of a city depending on the time of day. A fully-automated fleet could optimally use both sides of the road, on a 4-lane highway for instance by using 3 lanes in one direction in the morning and 3 in the other at night. Many rapidly-developing nations like India and the Philippines suffer from massive traffic jams as their infrastructure rapidly becomes overburdened by cars; it’s estimated that the Philippines loses up to $2.4B per day from traffic alone in wasted productivity time. Reducing that by even a factor of 50% would be an incredible boon to their economy and future development. Reduced congestion has secondary benefits as well, like reducing carbon emissions (fewer cars on the road at any given time, and more efficient behavior while they’re on the road — less “city” traffic).
And traffic reduction is just one more fairly obvious effect of an autonomous fleet of cars; there are a thousand more benefits, most of them likely unforeseen. You can make cars lighter or redesign the interior for safety purposes; a global taxi system could eliminate the need for parking or garages, transforming cities and suburbs and increasing efficiency of land use; biking and walking in cities may increase as automated vehicles are better equipped to avoid them, and so on. Current estimates suggest that the US alone could save between $700B and $2.2 trillion a year in the US alone, assuming 100% adoption of the technology. That’s a whopping 4-14% of GDP.
Of course, this doesn’t come without a cost; there are inevitable risks associated with this technology (malicious attacks on the transportation network, increased urban sprawl, etc.) But the biggest risk is that it never happens at all. The bureau of labor statistics suggests there are over 170,000 taxi drivers and chauffeurs and 10 times as many truck drivers in the US, and every single one of them will be put out of business. That’s an enormous demographic shift as over 1% of the labor force (currently ~155 million) is suddenly unemployed (although presumably many would be able to get jobs in other industries in the short term). Anyone familiar with cab agencies’ fights with Uber knows that these industries aren’t going to let the self-driving car take their jobs without a fight; they are going pursue every avenue to preserve their livelihoods. And while even $700B seems like a crazy number at a national level compared to a 1% bump in unemployment, at the local level you can bet it will start to skew in favor of the cabbies and truckers, especially in a time where preserving jobs and creating jobs are buzzwords on the campaign trail.
This is why I say we must come to some acceptable solution that allows the Googles of the world creating these innovative and transformative technologies to profit from them, yet preserves the quality of life of the millions that are displaced by it. We cannot allow ourselves to lose sight of the incredible potential global gains at the expense of the few; instead, we should apply a portion of those gains to mitigate the expenses. It’s the only way we move forward.