One of the main issue met by self-driving cars manufacturers is how to plan vehicles behaviour when crashing. Indeed, when the car detects several obstacles that can’t be avoided, it will up to the car’s artificial intelligence to chose who or what must be sacrificed. To help to make their minds on the subject,manufacturers have created an online survey to question people and analyse what the majority thinks.
Choosing who or what must be sacrificed when self-driving cars detects several obstacles that cannot be avoided has long been a major issue in the manufacturing of such vehicles. Far from being trivial, this question has forced some of the most important potential manufacturers to abandon the idea of developing self-driving cars.
Indeed, in 2016, Google, who was on the top 3 of the companies the most likely to achieve the manufacture of self-driving cars, announced that they had entirely abandoned the project. In 2017, Amazon did the same, focusing more on the manufacture of self-driving technology than on self-driving cars themselves. Apple and Uber also followed the same path, putting forward partnerships with big high-tech and car manufacturers rather than developing their own car branch.
In order to overcome this issue, self-driving cars manufacturers have created an online survey that will help them to teach self-driving vehicles’ artificial intelligence what to do when they meet several obstacles that cannot be avoided. Believe it or not, the majority of the 2 million people questioned thinks that old people and animals should be the ones to be hit when there is no other option.
Using an online survey, like the ones developed by Drag’n Survey, to make such a choice may seem quite irresponsible in a way, as even if a majority thinks so, it does not mean that it is the best thing to do. However, regarding the complexity of the situation, it is probably a good thing to ask people thoughts on it, rather than letting one or a group of persons making such an important choice.
In the survey, people were asked to imagine a situation where a deadly crash was going to occur and where the self-driving car had to choose between two sets of people. They must then tell which of those groups would die.
While the results were quite irregular between different groups of people, there were still a number of common belief between them and scientists found out 4 main elements :
– cars should first favour that fewer people die,
– older people rather than younger ones must be killed,
– humans must be favoured compared to animals
– cars should favour people who abide to law rather than people walking on the road for example.
These preferences seemed to be universally agreed upon and it is most likely that they will be integrated in self-driving cars’ artificial intelligence.
