What should a self-driving car do if it was faced with the choice between crashing five people on one side and just one person on the other? How should it react to the situation?
Well, let the car learn from the experts--the human beings themselves.
A research team in Germany put more than a hundred volunteers in a virtual reality simulation and had them make time-sensitive choices between crashing into children, adults, animals and inanimate objects to see how humans make decisions in a safe environment.
Researchers had volunteers wear an Oculus Rift headset and simulated driving a car down a street. They will be shown a pair of things in front of the car--from people and animals to inanimate objects. Each participant had to steer the car in either one of the things, sacrificing one to preserve the other.
After several test runs, researchers found mostly unsurprising results: participants were more willing to save a human than an animal and would run over an adult rather than a child.
"I think virtual reality is a breakthrough for empirical ethics, because without this there really is no way to reproduce in a controlled setting an experiment which really touches upon matters of life and death," lead author and PhD candidate Leon Sütfeld from the Osnabrück University told Motherboard.
"Studies show that there are vast differences between abstract situations and behavior in more realistic scenarios, so I think that VR will be a very useful, broadly used tool in the future."
However, humans as we are, we're also very conflicted. We have the natural instinct for self-preservation these VR simulations leave out. So, using algorithms to design automatons is really tricky just as how complicated each and every one of us makes decisions.
What do you think? Will this perspective be enough to be followed for a crucial technological breakthrough that could cost a life of a pedestrian inasmuch as the passengers over a bunch of squirrels?
Get weekly science updates in your inbox!