Search our Blog

Search our Blog

Tuesday, April 18, 2023

What Happened to Driverless Cars?

We were told self-driving cars would be here by now. Where are they?

We won’t see fully self-driving cars until 2030, and then only in the most expensive models, reports consultancy Accenture. How could that be? More than $100 billion has been put toward the technology, which has promised to advance consumer safety and convenience. Tesla founder Elon Musk has announced every year since 2017 that the means to create a true self-driving car will be available “next year.” 

But even the esteemed entrepreneur and billionaire has continually gotten it wrong. Why? It’s a lot harder to mimic common sense in software than anyone thought. 

Seniors Want Driverless Cars

The promise of autonomous vehicles is alluring. More Americans have been killed in car crashes than in all the country’s wars combined. We could potentially save lives while making car ownership unnecessary; why not summon a car from a nearby fleet to go to the store or make a cross-country trip?

Older people would no longer have to worry about failing eyesight or slower reaction times. If you could say, “Take me to Dr. Smith at Elmwood Heart Specialists” or “Drive me to the nearest grocery store,” you are going to get there. Seniors could look forward to a full lifetime of autonomous travel without depending on friends or family members. 

Edge Events

Most of the work of creating driverless vehicles has been done. They can handle changing traffic lights, turns, and other commonly encountered situations. What they have much more trouble with are the “edge events,” or unusual things that happen rarely overall, but are critical to safe driving. 

Edge events include a ball rolling into the street, followed by a child chasing it. It might be a tumbleweed blowing across a highway, a bicyclist who stops at an intersection, a sign that has fallen partway across the road, a dog running into traffic… and on and on. These are the situations that a human being can instantly process, but artificial intelligence (AI) struggles to interpret.

In one widely publicized incident, a woman was killed by a driverless car while walking her bike across the street. Autonomous vehicles have had trouble recognizing motorcycles and have rammed into the back of parked emergency vehicles, among other mishaps.

“I think if every car was a self-driving car, and the roads were all mapped perfectly, and there were no pedestrians around, then self-driving cars would be very reliable and trustworthy,” says Melanie Mitchell, computer scientist and professor of complexity at the Santa Fe Institute. “It’s just that there’s this whole ecosystem of humans and other cars driven by humans that AI just doesn’t have the intelligence yet to deal with.”

The Last Ten Percent

Humans can generalize from one situation to another, but if AI masters how to handle a situation, it cannot apply that knowledge to slightly changed circumstances. 

“It’s a challenge to try to give AI systems common sense, because we don’t even know how it works in ourselves,” says Mitchell.

There’s a saying in the tech world that once you have 90% of the work done, you only have 90% left to do. That last 10% of figuring out how to get something to work is always the most difficult. 

“It’s really, really hard,” said then-CEO of Google’s self-driving car company Waymo, John Krafcik, of the technology in 2018. “You don’t know what you don’t know until you’re really in there and trying to do things.”

Driverless Cars and Trucks Are on the Road

For now, companies are having success with driverless vehicles making deliveries, where passengers don’t have to be protected. Self-driving trucks are on the roads, often operating at night on highways, when there is less traffic and no pedestrians or bicycles to contend with.

GM’s Cruise is running a fully driverless robotaxi service in San Francisco, Austin, and Phoenix, but only after dark. The company was in the hole $1.4 billion last year alone, and it has yet to be seen how much more it’s willing to throw at the technology without turning a profit. 

Given how much work has been done in the field and the obvious benefits the technology could offer, it seems likely that we will see driverless cars on the road in the future. But older adults will have to hang in there several years longer before a more perfected technology ensures their rollout.