You may have heard the slosh of a collective eye roll last week from the Trough of Disillusionment crowd when Waymo said it had found that its system could have prevented the vast majority of traffic fatalities that occurred over the past decade in Chandler, Ariz., where it has been testing its robot-driven Chrysler Pacifica minivans.
Was Waymo’s analysis and its subsequent release self-serving? Sure, it was.
But it also serves the greater good in a couple of important ways.
It shows what can — and can’t yet — be done with automated driving, and it puts the technology’s focus rightly on its lifesaving potential.
But first, let’s look at Waymo’s analysis. Working with Arizona’s Department of Transportation and an independent third party, Waymo built mathematical models of 72 fatal crashes and ran simulations in which Waymo’s computer “driver” substituted for the vehicle that instigated the accident or others that had to respond.
To some extent, the game is tilted in Waymo’s favor. Assuming, as Waymo did, that the automated systems never malfunction, it’s designed to avoid a lot of human errors: Automated systems don’t get sleepy or angry or impatient or distracted or intoxicated. So right off the top, 52 of the scenarios went from tragic to uneventful.
It’s hard to appreciate the absence of a tragedy sometimes, but that’s what a thoughtful analysis can provide.
Because here’s the thing: When people die in a robot-driven car, there will be many outraged voices — some with agendas, others just scared. The idea that a person in a crashing car won’t be able to do anything to affect their fate other than buckle up and say their prayers — it’s a nightmare scenario.
There’s a kind of calculus that has to be worked through: How many lives can be saved? How many might be lost to imperfect technology or flawed algorithms or physical damage to vehicle sensors?
Before we as a society can turn over the steering wheel, there needs to be some comfort — at least on an intellectual level — that the technological solution is undeniably safer than the human driver.
And this is the kind of data that can spark those discussions. It’s a starting point for smarter policy debates, which hopefully will be happening in Congress this year under the leadership of Democratic Michigan Rep. Debbie Dingell.
Reducing deaths is a key goal of automated driving, along with preserving or restoring freedom for people who, for instance, can no longer see well enough to drive — or maybe never could. That has to be the centerpiece of the development, not the ability to start happy hour a little earlier or binge more shows on Netflix.
Deaths on U.S. roadways have stayed persistently high over the years as technological progress, such as adaptive cruise control, is overwhelmed by higher speeds and eyes looking down at a screen instead of looking down the road.
Just this month, the National Safety Council estimated that 42,060 people died on U.S. roads last year, up 8 percent from 2019. Apparently, while millions worked from home and stayed off the roads, many of those still out there decided to floor it, leaving themselves too little room for error.
As sad as America’s lack of progress has been, the figures from some other countries are just devastating.
China and India each lose more than 200,000 lives a year in road traffic incidents, according to the World Health Organization. They are the largest countries, with populations of more than 1 billion each, but they have fewer autos than the U.S. does.
Safety technology is expensive, and poorer countries suffer the most. Nigeria has fewer than two-thirds as many people as the U.S., but almost the exact same number of deaths. Ethiopia has one-third as many people as the U.S. but three-quarters the number of deaths.
As so many humans cope with deaths of loved ones due to the COVID-19 pandemic, and while the mourning process for all is restricted by safety protocols, it’s important to remember that a lot of the technology developed by this industry is about saving lives.
As it should be: It’s the right game to play.