According to Waymo, the Google division that operates self-driving robotaxis in several U. S. cities, their vehicles are involved in 80% fewer injury-causing crashes than human-driven cars. This was a surprise to me, as it may also have been to Kelsey Piper, who writes in a substack called The Argument that if we want to reduce the number of traffic fatalities in the U. S by 80%, all we have to do is switch to Waymos.
At the current rate of nearly 40,000 U. S. traffic fatalities a year, that would be saving 31,000 lives a year. But she cites a poll conducted by The Argument which found overall that only about 28% of respondents favored allowing self-driving cars in one's town or city, and 41% favored a ban, as Boston and other cities have considered doing.
Piper speculates that a general distaste for AI-powered things may be behind what looks like irrational opposition to self-driving cars. Some Waymos have even been attacked by modern-day Luddites. As she puts it, ". . . you can't vandalize ChatGPT, so anti-AI sentiment finds its expression in harrassing Waymos."
She realizes that anything like a wholesale move to self-driving cars would cause massive disruption to our present transportation system. She also acknowledges the tendency of government to make optional things mandatory. Right now, Waymo is one private company offering a specific service in a few carefully chosen markets such as Atlanta, Austin, San Francisco, and Phoenix. With the possible exception of Atlanta, these are all locations with plenty of sunshine and relatively few days of inclement weather. And the service in Atlanta only commenced last June, so Waymos in that city have not gone through a typical Atlanta winter, which almost always includes an ice storm, as I know from living there a couple of years. Ice may present serious challenges to Waymo's algorithms.
So there are practical limitations to the areas that Waymo can operate. As one of the commenters on Piper's article mentions, Waymo has somewhat cherry-picked markets in which it can compete while maintaining its very good safety record. Even if Boston decided to allow Waymos, I expect a long development period would delay its deployment as the company figured out how to deal with snow, ice, and slush on the former cowpaths that are now Boston streets. At least the Waymo cars probably wouldn't get lost as much as I did whenever I drove to Boston when I lived in New England, which was, well, pretty much every time I went there.
But there's those 31,000 lives. As Piper points out, that's more lives than are lost every year to homicides. Wouldn't it be nice if we could eliminate all homicides? Well, numerically, changing to self-driving cars would do that.
The psychology of why we tolerate such a lot of annual deaths to traffic accidents is interesting. People don't always act toward various risks in reasonable ways. The classic example is driving to the airport to fly somewhere. Unless you live within walking distance of the airport, you're going to drive or ride. And unless you take a Waymo there, you will ride in a human-operated car. While lots of people are afraid of flying much more than they are afraid of driving, the chances of dying in a plane crash are much lower than the chances of dying in a car wreck on the way to or from the airport.
I suspect that we have gotten so used to the 40,000 or so traffic deaths every year with some variation on the notion that it always happens to someone else, and if I were in the same situation that somebody else died in, I would have been clever enough to save myself. These are pure rationalizations, but the alternative is to experience a little squirt of adrenaline every time we buckle the seat belt and pull out of the garage.
Five or ten years ago, there was a lot of hype about how every new car would be self-driving within a few years. That obviously hasn't happened, for a variety of reasons. One factor is the expense per vehicle. Waymo doesn't advertise how much each of their vehicles cost, but various estimates say it's probably on the order of $160,000 to $300,000. This puts them in the super-luxury class, and explains why they aren't deployed in areas that will not generate pretty good revenue. Even with the carefully-chosen markets that Waymo has selected, it appears that they are not making a profit yet, which means the whole thing is still an elaborate experiment oriented toward some future situation that hasn't materialized.
Still, I will admit that if I could have a self-driving car that was absolutely trustworthy, a Level-5 one according to the SAE autonomous-vehicle rating system, one you could read or sleep in while the machine takes care of all driving tasks, and not pay a whole lot more for it than I'm paying now, I would at least consider it. But in my relatively small town, which would not provide enough revenue for Waymo under its present operating circumstances, that's not going to happen.
Now and then on my trips to Austin, I see a Waymo eerily coasting along with nobody inside, and once I saw two in a row. I will admit to having a flash of mischief when I saw one the first time, and wondering what would happen if you pulled in front of it and slammed on the brakes suddenly. Probably nothing bad. Fortunately, I was a passenger in the car I was in, not the driver, and so the experiment was never performed.
Unless something radical happens in the areas of AI, sensors, or legislation regarding the right to drive your own vehicle, it looks like Waymo and similar autonomous vehicle companies may not spread their wares much farther than they've already done. And that's too bad for the 40,000 or so people who die on roadways each year. Some ideas look good in theory, but when you start examining everything that would have to change to put them into practice, it just falls apart. And converting our fleet of vehicles to nearly 100% self-driving looks like another one of those nice ideas that has had an unfortunate encounter with reality.
Sources: A note in the online magazine The Dispatch referred me to Kelsey Piper's article "Please let the robots have this one," at https://www.theargumentmag.com/p/please-let-the-robots-have-this-one. I also referred to a Reddit item on the estimated cost of Waymo vehicles at https://www.reddit.com/r/SelfDrivingCars/comments/1g8vv7o/where_did_the_whole_talk_about_the_cost_of_waymo/, and the Wikipedia article "Waymo."
No comments:
Post a Comment