Across the globe, we can’t seem to agree on who self-driving cars should save in crashes
Self-driving cars promise to reduce the 40k+ car-crash deaths that occur every year in the US, but some deadly crashes will remain unavoidable: So when disaster strikes, how does a car choose who lives and dies?
Plenty of the brainy engineers have been asking that question — but, as a recent paper shows, they can’t agree on the answer.
Turns out, people across the globe can’t even decide whether cars should save babies or grandmas, not to mention passengers or pedestrians.
Autonomous auto ethics: Results may vary
MIT researchers asked 2.3m people from different cultures about the importance of several variables (young vs. old, rich vs. poor, law-abiding vs. jaywalking) in fatal crashes, and discovered ethics vary by country.
Respondents from individualistic cultures (France, US, UK) saved young people, while collectivists (China, Japan) chose old people. Citizens living under strong institutions (Finland, Japan) chose to hit jaywalkers more often than those with weak institutions (Nigeria, Pakistan).
Other results were harder to explain. Chinese respondents strongly chose passengers over pedestrians, while Japanese chose the exact opposite — posing challenges to auto-automakers trying to create consistent codes.
The pedestrian paradox
A few results are black and white: All cultures agree on prioritizing people over animals, and some suggestions — like saving rich people over poor people, as some respondents preferred — simply aren’t options.
Carmakers must decide whether to design and market cars differently based on the market’s expectations — selling passenger-prioritizing cars in China and pedestrian-prioritizing cars in Japan.