Now, we’re well into 2021, and I’m starting to get discouraged.
Then, I spoke to an expert I’ve known for years, who told me basically the same thing.
Actually, that’s not entirely true. Francesco Biondi did tell me we might see such a car, “but I think it will probably be toward the end of my lifetime.”
He is 33. I don’t stand a chance.
All those dreams I had of a car that would take me to doctor appointments in my old age, drop me off at the door, park itself, then come get me when I was done? Poof! Up in smoke, like so many semi-autonomous cars that run into things.
Biondi used to be an assistant professor at the University of Utah, where he conducted tests on volunteers who drove back and forth to Wendover in semi-autonomous cars. They drove one way manually and let the car handle much of the driving in the other direction. They wore contraptions that measured brain waves and had video recorders trained on them, which eventually demonstrated that people tend to lose focus when their cars automatically stay in their lanes and brake for traffic ahead.
Which is a problem, because those cars aren’t really ready to drive themselves.
In other words, it’s more dangerous to drive today’s semi-autonomous cars than to pilot that old, dusty beater that requires you to do everything — brake, accelerate, turn, switch on the radio and watch out for the other guy — all by yourself.
If your beater has a manual transmission, you don’t even have a hand free to text while driving.
Today, Biondi is a fully tenured assistant professor at the University of Windsor, in Canada. His window lets him look across the river toward Detroit, which once was the symbol of America’s booming automobile industry. He’s still very much involved in similar research.
“I think that the industry assumed it would be much easier,” he said about self-driving cars. The reason they were wrong is something that ought to make all of us feel proud:
“As much as we like to think that humans are not as smart as machines … that’s far from being the case. Even after years of testing, autonomous cars are unable to match the human driver’s performances. There are a lot of scenarios — weather conditions, road conditions — where these cars are not as smart as we want them to be.”
So, rogue machines aren’t going to take over the world any time soon. That’s good news, even though it won’t help me get to the doctor’s office as an old man.
And yet that message doesn’t seem to be widespread knowledge. At least, that wasn’t the case last April in Texas when two men decided to go for a ride in a Tesla, with one apparently in the back seat and the other in the front passenger seat — and no one behind the wheel.
The New York Times said their wives overheard the men talking about the car’s autopilot features as they left. They died when the car left the road on a steep turn.
Biondi said evidence shows many people who buy semi-autonomous cars overestimate what they can do. Dealerships sometimes oversell the products or leave out information about limitations.
Mims, of the Wall Street Journal, suggests some experts think the whole idea needs to be rethought. They want to approach the necessary artificial intelligence by building systems that learn the way human babies do. Others want to divide the problem into smaller chunks, developing systems that each handle a piece of the puzzle.
Biodi said he thinks commercial trucking shows more promise at the moment, especially if it involves dedicating lanes of traffic for that use only.
To me, it all sounds long and laborious and so … well … 20th century.
It also sounds like I’d better be prepared for that day when, as with my Dad and those before him, the kids discreetly take my keys from me and, in my decrepit state, I have to bum a ride to the doctor. What a bother.