Really great read and well written article! There are so many sides to this issue that still need to be flushed out, but ultimately autonomous cars will save millions of lives each year. As a computer science major, a lot of my professors have spoken about a huge obstacle that has to be overcome before autonomous vehicles are adopted by the masses, which is the countless moral dilemmas that software engineers face when coding the cars. For example, in a rare case where an autonomous vehicles's brakes go out and the car can either crash into another car, saving you (the driver) but killing a family of 5 inside of the other car, or crashing your car into a wall, killing only you but saving the family, which would you choose? These situations are all essentially variations of the classic "trolley problem" and there are a million different situations and variables that play into these no-win scenarios that could be faced by the car and it is all up to the programmers to decide how the car will react in these situations.
These dilemmas also bring up another problem of who is liable in these situations. Is the programmer, the car manufacturer or someone else liable for the crash? One infeasible way of solving the problem would be to allow the driver to select certain settings before the car hits the road, placing the blame on the driver. However, studies have shown that this would result in most cars being aggressive killing machines (they would mostly choose to kill the other car) since people think differently when it is their life on the line and not someone else's.
MIT has been doing extensive research into the subject and wrote a great article summarizing the problem. The article summed up: "Therein lies the paradox. People are in favor of cars that sacrifice the occupant to save other lives—as long they don’t have to drive one themselves... If fewer people buy self-driving cars because they are programmed to sacrifice their owners, then more people are likely to die because ordinary cars are involved in so many more accidents. The result is a Catch-22 situation." (https://www.technologyreview.com/s/542626/why-self-driving-cars-must-be-programmed-to-kill/)
Overall, companies manufacturing these autonomous cars need to be open and honest about how the cars will react before the situations arise in real life, or the adoption of self-driving cars will be greatly delayed. These issues cause people to lose track of the bigger picture, in that these cars will actually save millions of lives, directly impacting people's lives by making everyday driving much safer. Do you have any insights on the issue from working within one of these companies? If it isn't covered by the NDA, what are your companies current policies regarding these situations?
That situation has been the topic of many talks and debriefs since I have been here, but unfortunately I cannot say anything about our policies if it is not public information. I can say we haven't had any incidents yet in which we saw this first hand, so I am not sure how common this situation will actually be although I understand it is one of the biggest hurdles AVs will face. I can say our AVs are generally programmed to be more passive in any situation than any human will be. This is not always good but I feel this is how AVs will be tested until the technology is perfect.
Downvoting a post can decrease pending rewards and make it less visible. Common reasons:
Submit