I took my first ride in a self-driving car nearly a decade ago. Like most people who experience this brave new world, I felt a deep sense of awe that machines had mastered a skill that once belonged solely to humans.
Then I remembered I was a reporter.
Over the next 10 years, first at Wired Magazine, now at The New York Times, I covered the high-speed race to bring self-driving cars into the lives of everyday Americans. And during that time it became increasingly clear to me that although self-driving cars were shockingly nimble — indeed awe-inspiring — they could not yet match the power of the human brain. They still can’t.
My years of reporting culminated with an article my colleagues and I published last week about how driverless cars get help from humans.
Thanks to the multimedia talents of Jason Henry, Ben Laffin and Rebecca Lieberman, the article shows that although today’s robot taxis do not have drivers behind the steering wheels — some don’t even have steering wheels — they still lean on the good sense of people like you and me.
In April, my fellow reporter Yiwen Lu and I visited a command center in Foster City, Calif., operated by Zoox, a self-driving car company owned by the tech giant Amazon. Like other robot taxis, the company’s self-driving vehicles sometimes struggle to drive themselves, so they get help from human technicians sitting in a spacious room in the command center.
Sometimes, a technician lends a helping hand to a robot taxi hundreds of miles away.
If a Zoox car is unable to navigate a construction zone it has not seen before, for instance, a technician at the command center receives an alert. Then, using the computer mouse to draw a line across a digital road map on a computer screen, they give the car a new route to follow around the construction zone.
To truly understand this, you need to see it. Luckily, I work with people like Jason, Ben and Rebecca, who helped unspool the experience in visual detail for readers.
I had spent years talking to people about hidden human assistance. But I had not seen it with my own eyes until I walked into the Zoox command center. For years, companies like Waymo (owned by Alphabet, Google’s parent company) and Cruise (owned by General Motors) avoided any mention of the remote assistance they provided their self-driving cars.
That is just how things work in Silicon Valley. By creating the illusion of complete autonomy, companies can fuel interest in their technology and raise the billions of dollars they need to build a viable robot taxi service.
When I took my first self-driving ride in the mid-2010s, in a Waymo car, the hype surrounding the technology was approaching its peak. Across the industry, executives insisted autonomous vehicles would go mainstream by 2020. Elon Musk, Tesla’s chief executive, said it would happen even earlier.
Needless to say, those predictions did not hold up.
As companies tested their self-driving cars on city streets — and occasionally gave rides to industry insiders and reporters like me — they kept someone behind the wheel, just in case something went wrong. The companies called these people “safety drivers,” and made a point of noting that the safety drivers rarely intervened. Humans, the companies said, would soon be removed from the car.
At Tesla, Mr. Musk started selling cars with what he called “Full Self-Driving” technology. But if you bought and used the technology, you were legally required to keep your hands on the wheel and your eyes on the road, ready to take control at any moment. But he too insisted that in the near future, the human safety net would not be necessary.
My reporting told me otherwise. And as companies like Waymo and Cruise began to remove safety drivers from their cars, I began to see the limitations of the technology firsthand.
In the fall of 2022, the photographer Jason Henry and I took a ride in a Cruise car with no safety driver, and it got stuck in the middle of a sharp turn onto a San Francisco thoroughfare. We had to get out and call a new car.
A few weeks later, I spent a day in a Tesla with Chuck Cook, a man from Jacksonville, Fla., who was testing the technology. Full Self-Driving did not live up to its name: In a video, you can hear me yelp as the Tesla unexpectedly veers into a motel parking lot. Chuck had to take the wheel.
Now, Waymo-operated robot taxis — with no humans behind the wheels — are available to everyday Americans in places like San Francisco and Phoenix. Others, including Zoox, are not far behind. In both Foster City and Las Vegas, Zoox is testing a vehicle that doesn’t have a steering wheel. It doesn’t have a driver’s seat either. Or a gas pedal.
But don’t let that fool you. City streets can be chaotic places where the unexpected can happen at any moment. Machines are skillful. But they aren’t yet as skillful as humans when it comes to dealing with the unexpected. We can reason through these moments in ways that a self-driving car cannot.
These moments may be rare, but that is not the point. The point is: they happen.
That is why all robot taxi companies operate command centers like the one I visited in Foster City. Human assistance is essential.
What I have learned after more than 10 years of covering self-driving cars — and so many other forms of artificial intelligence — is that these technologies are not as powerful as they first seem. When we, the people, see a bit of human behavior in a machine, we tend to think, subconsciously, that it can do everything we can do.
But we should give ourselves more credit. Next to these machines, our brains are far more powerful than they might seem.
The post When Self-Driving Cars Don’t Actually Drive Themselves appeared first on New York Times.
