Alex Johnson
The technology behind driverless, “AI” cars seems to be advancing more quickly than people are comfortable with. Well, most people—some sleepy Tesla drivers are super-comfortable with computers taking the wheel.
Right now, a relative minority of people are comfortable ceding full control to their hardware, and even fewer can afford it. But the tech will get cheaper and more widespread. People will grow up with it, and eventually, fear of AI cars will seem antiquated.
It’s already easy to see how future generations will judge the very idea of humans driving cars. When you stop and really think about it, it already seems insane. People often cite air travel as being way safer than driving because it somehow makes us feel safer, even though many of us drive more than we fly. We have a blind spot when it comes to driving.
We let people drive for a century or two because we made cars before we could make AI. If we could (mostly) replace horses as a premier means of land travel after nearly 6,000 years, replacing manually driven cars as the standard will be nothing. Progress will march on, and all but the old-fashioned hobbyists will forget how to drive. But we’ll also forget that we ever cared about knowing how to drive.
Car manufacturers have been trying to make cars that drive themselves since the early 20th century. At first, they used things like magnets and radio control. It wasn’t AI, but it betrayed an innate human desire: to sleep while driving.
Now, here we are in the 21st century, and the tech race toward driverless car technology is on. BMW has been working on driverless tech since 2005 or so. In 2010, a driverless Audi TTS was tested at near-race speed, and GM made its urban Electric Networked Vehicle (EN-V) the following year. The Volkswagen Temporary Auto Pilot System, which began testing in 2012, can drive itself at 80 mph (on the highway—no Bourne Identity car chases in the city while you nap, yet).
And of course, there’s the Tesla, which is currently the most popular driverless car being abused by early adopters today (on the internet, anyway). I won’t be impressed until I see a video of someone taking a nap while off-roading in the Audi AI: Trail, which will have drones for headlights.
And it’s not just car companies jumping on the bandwagon. Google, for example, created the now-stand-alone subsidiary Waymo, which is working on a self-driving taxi service. It’ll be like current ride-share options, minus the part where a human makes extra money to pay rent. Apple is also working on something, presumably a Bluetooth-only car with no audio-in jack.
But, as I write this, we still don’t have fully driverless cars. They tend to be highway-only and fall under the umbrella of “assisting” drivers, rather than replacing them—at least in terms of what the public has access to. Elon Musk says the Tesla will be there soon, but there are a lot of limitations on AI right now that call for some skepticism on that account. So far, the only real public beta testers for AI cars without a human fail-safe—the real pioneers—are those willing to fall asleep on the highway while their car goes 70 mph.
The Possible Future of AI Cars
For me, the impact AI will have on how we drive will be determined by two things: the advancement of the tech and our willingness to adopt it. Whenever I try to imagine how quickly things can change, I like to think about the Wright Brothers’ first powered flight. In 1903, they managed to keep a primitive airplane going for a few hundred feet at about 6.8 mph and about 10 feet off the ground. Then, like 65 years later, NASA landed some fellas on the moon.
And that was before the technology revolution in which we currently find ourselves. Culture also moves faster now (in some ways), so both the tech and our acceptance of it likely will happen more quickly than we think.
There will be innovations I can’t fathom, but it seems a safe bet we’ll have fully driverless cars in the not-too-distant future—if only so the industry can fire truck and cab drivers.
We already accept AI assistance in planes and trains, and we’re happy enough to cede control of our safety to those machines. Hell, it might even make us feel better to know that a computer is in charge, rather than a sleepy person. Many of us already trust machines more than other people, but the trick with AI in cars is you have to trust a machine more than yourself.
What It Means for Drivers
Once we finally give it up and hand the keys over to our cars, most folks are going to lose (or rather, not gain) a skill set. Think of manual versus automatic transmissions: automatic became the norm, so a lot of people (like me) never had to learn how to drive a stick.
Even now, a lot of people either never learned to drive or choose not to—especially in urban areas. Many New York City dwellers forego a car (and the nutty parking and traffic). And many who were born there might have never taken driver’s education.
In Copenhagen, concerted efforts have yielded a city with more cyclists than drivers. Cities with weaker infrastructures also provide an incentive to go car-free. I have felt a deep pang of envy as I watch a cyclist pass by, while I sit in bumper-to-bumper, Boston traffic. When you consider 68 percent of the world’s population will likely live in cities by 2050, it’s not just AI cars that will lead us away from driving, but where we live.
People who really dig driving will still learn how to drive, just as some still learn to drive a stick shift. But for those of us who just need to get from point A to point B, we’ll only learn what we need to know to pass the driver’s test. For that future test, you might only need to know how to pull the car over and hit a button to call for help because the AI failed. Who knows?
What I do know is how easy it is to forget the skills you used to have. I remember printing out MapQuest directions to different cities back in the early 2000s. It was easier than reading a highway map, but MapQuest still demanded some things from you. You had to have a general understanding of where you were going before you hit the road, and you had to judge how many miles you’d gone. You also had to pay attention so you didn’t miss your exit—paper doesn’t recalculate your location and provide a new route.
In 2019, I turn on my GPS even if I know where I’m going, just to have an idea of when I’ll get there. The moment it miscalculates or loses a connection with the satellite system, my anxiety levels spike. My brain’s been spoiled by the tech.
Even smaller luxuries, like side-view cameras, impact well-trained skills. When I first got a car with a side-view camera, I didn’t use it. It was baked in my brain that when I merged onto the highway, I had to turn my head and check the blind spot. I knew the camera would do that for me, but it took a while to retrain myself.
Then, I got an older car without side-view cameras and had to retrain myself to twist my head and look for the blind spot again. If you don’t use it, you lose it—or, at least, I do. Even if the tech becomes extremely reliable, it’s concerning that the roads might someday be filled with people who don’t know how to operate their vehicles. But I’m a product of my time and environment. Future generations will have a different idea of what’s possible and good.
People Aren’t Good at Stuff
There’s one particular thing we modern folk ignore so we can get on with our commutes: our greater-ape brains have inadequate reaction times for stimuli at driving speeds. Also, a lot of us are real jerks.
In that spirit, I think we should ultimately hand the keys over to the computers. We can’t react quickly enough to avoid all accidents. An AI, however, can be designed explicitly for that purpose.
At present, the public reaction to driverless cars tends to focus on its failures. When a self-driving car operating in a beta testing zone kills a pedestrian, it’s news. And it’s bigger news than a human driver who kills a pedestrian, or a driverless car that works perfectly well and harms no one. This makes sense, though, because AI cars are new, and people want to know if they’re perfectly safe.
But the real question shouldn’t be if AI cars are perfectly safe, but rather, if they’re safer than those driven by people. In the U.S. alone, nearly 40,000 people are killed annually in road accidents, and another 2.35 million are injured or disabled.
Will advanced AI cars cause or fail to avoid accidents some of the time? Sure; but as long as they do it less than humans, the benefits are too great to ignore.
Sometimes, It’s Good to Forget
I imagine the shift will happen more quickly than the leap from powered gliding to lunar landings. And my assumption is we’ll forget a lot of what we know about driving. Will that be a problem? Eh, I don’t think so. I don’t know how to ride a horse, and that causes me no problems in my daily life. But, not so long ago, if you didn’t know how to ride a horse, it might’ve made you a useless kind of person.
Things become obsolete. People move on, unless they pursue the old way of doing things as a specific interest. Some folks still ride horses or flint knap their own tools. Someday, people might go to Ye Olde Closed Track, sign a waiver, and try to pilot an antique “dumb” car.
The broader picture for society, at least from a road-safety perspective, is driverless cars will almost certainly lower annual traffic fatalities. Frankly, those sleepy Tesla drivers already (and irresponsibly) show that the tech can do a bit more than what it’s marketed for. Plus, humans are really bad at driving. There’s plenty of room for improvement, and we see it with the developing tech.
As for the impact on our skills and culture? Yeah, it’ll change those. But those things have always been fluid and change is inexorable.
Future generations simply won’t care that their dead ancestors liked to drive.
