Waymo's data from nearly 100 million driverless miles shows 91% fewer serious crashes and 80% fewer injury crashes than human drivers. TJ Walker argues autonomous vehicles could prevent roughly 1.2 million of the 1.3 million annual global traffic deaths, though he acknowledges concerns about job displacement and tech power.
Questions this video answers
Nearly 100 million fully driverless miles in four U.S. cities.
91% fewer serious-injury-or-worse crashes compared with human drivers on the same roads.
80% fewer injury crashes overall compared with human drivers on the same roads.
96% lower rate of injury crashes at intersections compared with human drivers.
About 1.3 million people die in traffic crashes every year globally.
Chapters
Self driving cars, Self driving cars and the lives we're leaving on the table. According to the Times, neurosurgeon Jonathan Slotkin argues that autonomous cars aren't just a tech toy, they're a public health intervention. Waymo just released safety data from nearly 100 million fully driverless miles in four US cities. Compared with human drivers on the same roads, its cars had dramatically and I mean dramatically fewer crashes, 91% fewer serious injury or worse crashes, 80% fewer injury crashes overall, and a 96% lower rate of injury crashes at intersection. Globally, about 1.3 million people die in traffic crashes every year. If we could impose Waymo level fatality rates on the entire world fleet, we prevent roughly 1.2 million of those deaths annually. More than 3,000 people every single day. In the United States alone, crashes kill more than 39,000 people every year. They are a leading cause of death for children and young adults and cost the nation more than a trillion dollars in economic and quality of life damage. Slotkin's message is blunt. Keeping humans behind the wheel is the riskier experiment. Okay, so what's the bottom line here? As much as I, a lot of people like to pile on our tech overlords having too much power over our lives, this is one case where we need to give more power to the tech overlords because human beings are horrible drivers. There, I said it. 1.3 million people die a year around the globe because of car crashes. And autonomous vehicles, they get more headlines when they crash, but they save lives dramatically. More than a million lives year would be saved if we just got rid of human drivers. Let's dig a little deeper here because this is a situation that cuts in a lot of different ways. I know many people who, they're pro vaccine, they're pro eating, right, they're pro minimizing risk in every aspect of life. And you say, hey, you want to go in this driverless car? Oh no, that's crazy. It's too risky. Well, it's not too risky compared to a human driver. The problem is we humans are awful. We're easily distracted, we hear things, we remember things. We're on the phone, we're texting when we're driving. Some of us are drunk when we drive. So almost anything is safer than getting in a car driven by a human being. Even if it's, believe it or not, you or me, we don't like to admit that, but it is simply true. Now here's some interesting statistics. Waymo has never rear ended another vehicle. And an injury Level created injury. Yes, there have been mistakes and there have been deaths. But why, if you're a rational person, would you say, okay, one death here from an autonomous vehicle is so much worse than 100 deaths over here by human beings. If you purport to be someone concerned about human life, as I think I am and you are too, why would you say it's better for people to die simply because technology wasn't involved? Now, I'm all in favor of people scrutinizing technology in their lives. And if you want to eliminate risk, how about getting your preteens and even teenagers off cell phones and social media? I think there's a lot of evidence that big technology is causing tremendous problems, divisions in our country and the world, civil unrest, income inequality. There's a whole lot of problems that social media and technology's cost. You could certainly make a case that people like Elon Musk have too much power when they can write a check for a quarter of a billion dollars and then essentially get to run the government like a play toy for a few months. So look, please do not put me in the category of, oh, TJ's just a big apologist. Anything Big Tech wants, anything Waymo wants, anything Google wants, give it to him. No, I'm not in that camp. I'm here as a sane, sensible centrist and I'm trying to really look at everything on a fact by fact, case by case basis. And when I look at a case for autonomous vehicles, I think even if they didn't get any better, it would be irrational for any human being who has a choice between an autonomous vehicle and a human driver to get into a human driven car. Now, just the past week I was with family and friends and one of them said, oh, a friend of mine's father in law was in an autonomous car and it went through a blocked off area in the street on a ramp and another one and he finally had to pull it stop and it would have gone off the cliff. Okay, but that very same person likely has a handful or a dozen extended family and friends died in human driven car crash over their lifetime. And yet somehow, because it's so common, it seems to not count. So I think it's time we all kind of grow up and look at the facts and not let various biases influence us. So I want to take a little time here today and talk about why you feel nervous and why you're not a crazy person if you're anxious about getting in a Waymo car. I did get into a Waymo autonomous driven vehicle a little more than a year ago. And hey, I'm a human being. I was a little bit nervous the first few minutes, but then I noticed it was smooth. It stopped and started. It kept a polite distance from other cars. It used turn indicators to change lanes. It did everything by the book. And I was delivered into a couple of locations safely, smoothly and efficiently. So let's look at some of the biases that I think influence all of us when it comes to technology. And just things like this that seem dangerous, more dangerous than what we're already used to, when in fact we've just become numb to the idea of that. Roughly. I hate using this comparison, but it's the most apt thing possible. There's essentially a 911 worth of victims, meaning more than 3,000 people a day die every day around the globe from car crashes. If we had someone from outer space bombing a giant skyscraper and 3,000 people a day died around the world somewhere, I think the whole world would be up in arms. We'd say anything to stop that. Any technology will do anything to stop 3,000 people a day dying. And yet, because people die in car crashes one at a time here and there, dispersed throughout the world, just seems normal. Nothing to look at here. So I want to sort of catalog some of these biases that I think are important to realize. There's the availability heuristic. People judge risk by how easily vivid examples come to mind. So a single sensation of a self driving crash, it feels more dangerous because we can see that, we can imagine it, it's, it's visible to us probably on the local news if it happened, or the national news. Whereas somebody just dying in a hit and run or a drunk driving incident or a head on collision doesn't even make the news in most towns because it's so common. The next kind of bias, there's a negativity bias we all face. Our brains give more weight to bad outcomes than good ones. So when we think of an autonomous driven vehicle, we instantly think about, oh, could be risky, we could die and crash. And what if it goes into something and I can't put my foot on the break. That's a negativity bias. So rare failures of autonomous cars loom larger than the far more common lives that it could save. We don't visualize the 1 million lives that could be saved next year if we could magically convert all cars into autonomous driving cars. There's the salience bias. Dramatic, unusual events grab our attention. A weird robotaxi mistake gets a Lot of attention. It feels important because normal human crashes are ordinary, boring. We don't pay attention. There's also something known as a status quo bias. People prefer the existing ways of doing things. So familiar human driving feels normal even though it produces massive death. Again, even here in the United States, people driving newer, safer vehicles, more safety laws here than most places. 39,000 people die a year in car crashes. Most of them wouldn't die if we had autonomous vehicles everywhere. There's the ambiguity, aversion. When outcomes feel uncertain or poorly understood, like a new AI system, people instinctively avoid them, even if the data shows lower risk. And let me be clear, there's a lot of areas where AI is trying to muscle into your life, where I think you should be fearful, you should say averted. Look, we don't really know what this is going to do. Letting our child go from playing outdoors three hours a day the way kids used to, to hunched over looking at TikTok five hours a day. I think there's good reason to fear that. But there's already evidence that these autonomous cars per mile save lives. There's also something known as omission bias. Many are troubled by harm caused by an active choice riding in a robo taxi than by an equal or greater harm from inaction, such as just continuing to drive a normal car. You know, think of it this way. I mean, I. Look, we're all humans. We all have irrational fears. I fear jumping out of an airplane and skydiving. At some point, I may do it, but there's still a sense of fear there where if I were a rational human being, I would realize driving to the airport to get in the plane is actually much more dangerous than jumping out of the plane. It's why people fear going to a beach because someone got attacked by a shark a thousand miles. Oh, what about shark? We fear that shark attack because it's rare, because it's visual, because it's extreme, because we just saw it in the news. Whereas we don't worry about the drive to the beach or plan B, the drive to the park, even though that's much, much deadlier. So put me in the camp of saying we've all got to experiment and adapt to these driverless cars as quickly as possible. Yes, I know it's wildly disruptive to our economy. Here in The United States, 2 to 3 million car drivers, Uber drivers, taxi drivers, truck drivers, are going to be put out of business. I'm in favor of retraining them. I'm in favor of subsidies. I'm in favor of giving them new education. Let's help them any way we can. But I don't want to help them if it means 39,000 people have to die needlessly in the United States every year. Now we're not. No one's going to wave a magic wand and replace all cars driven by people. I realize that. But Whether it takes 20 years or 30 years, I think it's going to get to the point where driving a car just for fun because you like it is going to be seen as eccentric, as using a horse to get someplace. Yeah. People like to ride horses from time to time, not usually as a mode of transportation. It's going to be seen as sort of an eccentric hobby. And meanwhile, we're going to see a lot of lives saved.
