⏩ Let's make self-driving cars happen faster!
A Q&A on autonomous vehicles with policy analyst Jordan McGillis
Quote of the Issue
“The current wave of advances in artificial intelligence doesn’t actually bring us intelligence but instead a critical component of intelligence: prediction.” - Ajay Agrawal, Joshua Gans, and Avi Goldfarb, Prediction Machines: The Simple Economics of Artificial Intelligence
⏩ Let's make self-driving cars happen faster!
🚗 A Q&A on autonomous vehicles with policy analyst Jordan McGillis
Earlier this month, California regulators agreed to the expansion of driverless taxi services in San Francisco, allowing companies Cruise and Waymo to offer paid rides anytime during the day throughout the city. It was the latest in several incremental but important advances for autonomous vehicles. Previously this year, Waymo expanded its service to the greater Phoenix area, while Cruise has expanded its robotaxi service to Houston and Dallas. Waymo also recently announced it will expand its autonomous rideshares to Austin, Texas, its fourth city.
But a faster rollout would be great given the huge potential upside. In a recent essay, I highlighted a thorough Manhattan Institute report on autonomous vehicles, “Autonomous Now: Why We Need Self-Driving Technology and How We Can Get It Faster.” It’s an in-depth treatment of self-driving technology, its implications for the economy in both ride-hailing and freight trucking, and the policies local, state, and federal lawmakers should consider as this technology develops. In short, there’s a lot of potential for AVs, which left me wondering why I hear so much skepticism, on the one hand, and hysteria, on the other. So I took my lingering questions to the author, Manhattan Institute Paulson Policy Analyst Jordan McGillis, who didn’t disappoint.
1/ You write that autonomous vehicles are "already safer than human drivers in many circumstances ... [and] will soon surpass human driving competency in almost all circumstances." Why, then, are we not hearing much more optimism about AVs? Why all the pessimism?
I think the first thing to mention is that the operation of the multi-ton moving metal machines we call cars is something that we all know comes with risk. An out-of-control car is obviously very dangerous, both to its occupants and to those outside of it. In the first half of the 20th century, many American cities witnessed mass demonstrations against cars due to the norm disruptions they caused and the grisly physical harms they inflicted. The scholar Peter Norton has a book that documents this history titled Fighting Traffic: The Dawn of the Motor Age in the American City.
In the postwar era, though, cars became a central part of the suburbanized American Dream, and outright opposition to their presence became a fringe position. Whether we attribute that shift to policy or to organic preferences, I think it’s safe to say most people are pretty comfortable now in 2023 with cars being the default way of getting around most places in America. We have a new status quo, yet, here come AVs to disrupt things again.
My position is that AVs can mitigate some of the risk that cars pose, while continuing to deliver the point-to-point transportation advantage that they confer over alternatives like public transit. But using autonomous driving systems to control vehicles is admittedly a radical change. It’s not surprising to me that most Americans are wary. We deal with a lot of personal consumer technology these days and, although it’s amazing most of the time, it’s not perfect. It’s logical to infer that while a smartphone’s glitching isn’t a big deal, if that same kind of thing happens with, again, a multi-ton moving metal machine very bad things will happen. In the past year a lot of us have interacted more regularly with artificial intelligence through things like Chat-GPT and Midjourney and we know that AIs get things wrong sometimes. Moreover, though the average person probably wouldn’t use the term “edge cases”, that person knows that a lot of weird things can come up on the road and they doubt that autonomous systems can handle the unexpected. I want to emphasize that these concerns aren’t crazy.
That said, our comfort with the status quo blinds us to the daily risks that human drivers pose. These statistics are well-worn by now, but they deserve repeating: 40,000 people are being killed each year on American roads and another 3 million are being sent to emergency rooms. (Globally, some estimates suggest a million people are killed on the road annually.) Overwhelmingly, human error is the proximate cause of these casualties. But we’re used to this. And we’re used to driving every day. Most of us think we’re good at it; it’s those other guys who are causing all the harm.
I would argue we’re not that great. We succumb to distraction, fatigue, and so on and so forth. We also underestimate the volume of experience that autonomous driving systems are trained on. They have access, actually, to far more driving data than a normal human has.
There’s another, almost polar opposite, case against AVs despite their stellar, growing performance record. The second case comes from urbanists who want to overhaul our cities totally to exclude or severely limit car use. They think AVs entrench dangerous car culture. I argue these urbanists are incorrect, even though I share some of their inclinations against existing car norms. Among others, one big empirical thing they miss in terms of safety is that ride-hailing (like Uber and Lyft) has already been shown to make roads safer by enticing people who have been drinking alcohol into safe transportation rather than driving themselves. When autonomous driving systems enable lower fares than the ride-hail options we have today, that effect will be even greater.
Keep reading with a 7-day free trial
Subscribe to Faster, Please! to keep reading this post and get 7 days of free access to the full post archives.