Moral Machines - TOUGH QUESTIONS ABOUT SELF-DRIVING CARS

Have you ever heard of the trolley problem? It is a set of imaginary scenarios used by ethicists— people who ask questions about right and wrong.

- Fiona Young-Brown

Here’s one scenario. Say you are driving a car down a curvy mountain road when you suddenly meet a flock of sheep. You don’t have time to stop. You could swerve into the mountain (and probably die). You could swerve in the other direction, off the side of the mountain (and definitely die). Or you could plow into the sheep (which will probably kill some sheep and damage your car, but you will likely be OK). Yikes. What do you do?

What if it is not a flock of sheep but a pair of bicyclists? Would your answer be different now?

Now imagine you are in heavy traffic and your car’s brakes start to fail. If you keep going straight, you will go into the back of a semi-truck and probably die. But if you swerve to avoid the truck, you risk hitting an elderly woman or a group of children. What do you do?

These are extreme choices. But every day, drivers around the world have to make split-second decisions to avoid accidents. Often, they react without even having time to think. But what would a driverless car do? Self-driving cars are programmed to deal with all sorts of situations, from understanding traffic rules to planning the easiest route that avoids construction. These programming decisions are straightforward. Ethics, on the other hand, is not. How do you program a car to decide what to do in an accident? How do you choose which life is more valuable? And who gets to choose? Should it be the car owner? The manufacturer? The government? This is a problem that ethicists, lawyers, and carmakers are all wondering about.

The Moral Machine Test