We’ve watched technology change drastically in the last few decades. Just 20 years ago, cellphones were just starting to become a thing, and now nearly everyone has a phone in their pocket. In fact, most of them are equipped with computers more powerful than the ones that helped Apollo 11 land on the moon. Until recently, getting behind the wheel of a car meant you had to drive it, but now, you can pick a car that can drive itself under certain circumstances.
While driverless cars are getting more and more popular, our driver-free future might be a little further away than we thought.
You can predict how a computer is going to react to any given scenario. It is programmed with a specific set of parameters, and it can’t act outside of them.
Human beings, on the other hand, are inherently unpredictable. You expect them to go left, but they go right. When they start acting randomly — especially on the road — the programming inside a driverless car can’t react in time. The result is self-driving car accidents, which give all driverless cars a bad rep.
Self-driving cars aren’t all that affordable right now either, putting them out of reach for the average driver. Currently, having a car with self-driving technology installed adds between $7,000 and $10,000 to the car’s sticker price, taking an already pricy car and making it even more expensive. When the average American can only afford to spend a little over $20,000 on a new car, that puts most self-driving cars well out of reach.
Another problem many in the car industry are facing is how to insure a driverless car. Until these cars manage to eliminate accidents, insurance will be necessary. If the driver isn’t the one responsible for the car accident, how do insurance companies need to respond?
One solution is to shift the responsibility from the driver to the car manufacturer. If the driver isn’t expected to actually drive the car, then they can’t reasonably be expected to take responsibility for the accident, right?
This is something the government will likely have to mediate before self-driving car insurance becomes mainstream. When it comes down to it, though, it will be up to insurance companies to make the final decision on how to insure driverless vehicles.
The Trolley Problem
One of the biggest hurdles self-driving cars are facing is the trolley problem. It’s an ethics question — if you’ve got one person tied to trolley tracks, and a group of people tied to another set of tracks, and you can control which track the trolley will travel down, will you kill one person to save the group?
This is a question posed to humans, who have the capability to make a decision like this. While neither outcome is ideal, the general consensus is that the needs of the many outweigh the needs of the one, so you kill the single person to save the group.
How do you teach that to a computer? Do you program a self-driving car to take all steps necessary to protect the driver and passengers, regardless of the people around the car? Do you try to teach it the ethics necessary to make a decision on something like the Trolley Problem?
It’s a question Uber has had to face recently, when one of its self-driving cars in Arizona hit and killed a pedestrian who was crossing the road in the dark. These cars may never be so sophisticated that they can determine who they should kill in real time in the event of an accident, but that could be one thing car manufacturers are working toward.
Driverless cars may be further away than we like to think, but they’re still coming. Many countries are working toward getting driverless cars on the road in the next decade, so we will have to wait and see what they do to make these vehicles safe enough to share space on the road with human drivers.