The trolley problem
This situation got us thinking again about a famous cognitive science experiment designed in the 1960s by the philosopher Philippa Foot: the trolley problem (opens in new window). It goes like this: through a thought experiment or some technical contrivance, you find yourself at the controls of a trolley whose brakes have failed. The trolley is racing toward a group of 5 people walking on the track, who won’t survive the impact. But you can divert it onto another track by pressing a button. On that track stands a single person. You face a moral dilemma. Do you make the conscious, deliberate choice to kill one person to save 5? Or take no responsibility and let the 5 die?
Two approaches then come into conflict: a so-called utilitarian approach, adopted by the large majority of subjects (90% in general), which will preserve the maximum number of human lives. Diverted, the trolley ends up on the track where the single person stands. This decision falls under what philosophy calls consequentialism. Any action becomes measurable against its moral consequences.
The other approach is to do nothing, on the grounds that you cannot put a number on the value of a life, and that the choice therefore cannot rest with the person at the controls. It rests instead with the designers. Lessons have to be drawn so that the situation never happens again (and that’s usually where we come in!).
Self-driving cars and moral decision-making
This brings us back to the question of self-driving cars: what ethics should designers build into their vehicle? In the event of a potential accident, should the car save its driver or several pedestrians? Jean-François Bonnefon of the CNRS CLLE laboratory (with whom we have worked in the past), together with Iyad Rahwan of the Massachusetts Institute of Technology (MIT) and Azim Shariff of the University of Oregon, investigated this question by putting it to a panel of 1,000 people.
It turns out that, as in the trolley problem, people would rather sacrifice the vehicle and its driver than the pedestrians… as long as they’re not the one driving. Morality counts for little against the survival instinct! So the vehicle would end up making a decision that goes against its owner’s wishes? To explore these questions in more depth, discover our study on driver experience and self-driving vehicles.
Humans and their decision biases
But would it really be any better to leave the decision to humans, those great specialists in irrational choices? We know, for instance, that the smell of warm croissants makes them more generous and more inclined to help others. We also know that under stress, humans release cortisol, which dulls their ability to think rationally and pushes them to look out for nothing but their own survival. We regularly study this kind of human behaviour using techniques such as eye-tracking.
Conclusion on self-driving cars
All that’s left is to expose self-driving cars to the smell of warm croissants… The question is still far from settled! To design effective and ethical solutions in this field, we use methods such as design thinking and the creation of prototypes, which make it possible to test critical scenarios. If the topic of artificial intelligence interests you, we point you to another blog article.
References
BONNEFON, Jean-François, SHARIFF, Azim, and RAHWAN, Iyad. Autonomous Vehicles Need Experimental Ethics: Are We Ready for Utilitarian Cars?. arXiv preprint arXiv:1510.03346, 2015.
R. A. Baron, “The Sweet Smell Of… Helping: Effects of Pleasant Ambient Fragrance on Prosocial Behavior in Shopping Malls”, Personality and Social Psychology Bulletin, 23, 1997, p. 498-503
