This past week, the US Department of Transportation released safety guidelines to auto manufacturers regarding the production of autonomous and semiautonomous cars. This is broken down by the New York Times here. One of the points that caught my eye was the “Ethical Considerations” section, and an example it gave, “should a car be programmed to better protect its occupants or other drivers in a crash?” In this situation, the autonomous car supposedly would have to be able to calculate the worth of the people inside and outside the car, and then based on those values, make a decision on whom to better protect. Are the lives of the 5 passengers inside the car more valuable than the 6 pedestrians on the sidewalk? Could this also become a point of competition between manufacturers? It’s reasonable that one would prefer to buy from the manufacturer that promises to always put the passenger first. It seems that most people are in favor of the car making the decision that minimizes the loss of life, but it also makes people wary of buying an autonomous car for themselves.
Everyday economics from the University of Puget Sound
This issue of autonomous cars making life-and-death judgement calls brings to mind the classic philosophy ‘Trolley Problem’ (which I always thought used a rather silly example) where individuals are asked if they would push one portly individual in front of a speeding trolley (or some equivalent variation of killing one person) to stop it from crashing, killing more people.
Interestingly, MITs Media Lab created a site called Moral Machine (moralmachine.mit.edu) to present people and survey them regarding just such issues.