September 1, 2023
A year ago, over at AIER, I noted that preventing self-driving cars when they are less than perfectly safe may cost lives.
I wrote:
“Discussions of how regulation could “get in front of” self-driving cars are therefore incomplete, and ultimately, may cost lives. According to the National Highway Traffic Safety Administration, over 42,000 people perished on U.S. roads in 2021. What that implies is that self-driving cars would be an improvement if, with autonomous vehicles widely prevalent, “only” 41,000 people were to perish in car accidents.
To put this even more starkly, were those numbers accurate, it would imply that every year regulators delay because driverless cars are not yet perfectly safe, they would be killing a thousand people on net.”
It’s the Nirvana Fallacy behind the wheel.
AI expert Timothy Lee amasses evidence that (at least some) self-driving cars have already surmounted the magic threshold. He writes:
“But we actually do know a fair amount about the safety of driverless taxis. Waymo and Cruise have driven a combined total of 8 million driverless miles, including more than 4 million in San Francisco since the start of 2023.
And because California law requires self-driving companies to report every significant crash, we know a lot about how they’ve performed.
For this story, I read through every crash report Waymo and Cruise filed in California this year, as well as reports each company filed about the performance of their driverless vehicles (with no safety drivers) prior to 2023. In total, the two companies reported 102 crashes involving driverless vehicles. That may sound like a lot, but they happened over roughly 6 million miles of driving. That works out to one crash for every 60,000 miles, which is about five years of driving for a typical human motorist.
These were overwhelmingly low-speed collisions that did not pose a serious safety risk. A large majority appeared to be the fault of the other driver. This was particularly true for Waymo, whose biggest driving errors included side-swiping an abandoned shopping cart and clipping a parked car’s bumper while pulling over to the curb.”
Also interesting is that, per Lee, we know little about how safe human drivers actually are.
Lee, again:
“We know very little about the safety of our roads,” the legal scholar Bryant Walker Smith told me. “If we're looking at just crashes, given how little information is carefully collected and studied, we don't have any sense of the circumstances of these low-level crashes.
Not all crashes—even serious ones—are reported to the police.”
Read the whole thing here.
As Joel Mokyr observes:
“…the surprising thing is not that there has been so much resistance to technological progress, but that humanity has actually been able to overcome most of it. For most of human history…the heavy hand of neophobia has suppressed intellectual innovators of all kinds, condemning many past societies to a fate of virtual technological stasis.”
If by “neophobia” we mean special interests protecting rents, I’m on board.