Obviously not. The pedestrian is responsible of his movements and can avoid the car. On the other hand the passenger is prevented of having any action. Therefore he should be protected at all cost by the car.
Now if the driverless car is empty, then of course, it should gladly destroy itself to avoid hurting human pedestrians, and even any animal or robot.
If the driverless car transports a human or animal passenger, then it should avoid hurting them, even if that means hurting external humans, animals or robots.
But if the driverless car transports a robot, then it should avoid hurting it, even it that means hurting an external robot. But it may hurt its passenger robot to avoid hurting an external human or animal.
Now, assuming some mean of communication between the driverless car and the external robot, and the passenger robot, the driverless car could, if it was absolutely impossible to find any combined action of the external robot and the driverless car avoiding to hurt both of them, then a solution that would hurt one or the other might be found, depending on the criterial of "importance" of each robots. For example, if one or both of them contain information or transports materials critical to save the life of humans or animals in an immediate situation (the reason why one, the other or both robots are moving), then the choice should be to keep moving the one with the best impact.
For example, the robotcar drives a robot transporting the antitoxin that will neutralize a biobomb that could kill 100,000 humans in ten minutes, while the external robot transports the bronchodilator that will save an asthmatic child, and there's no way the robocar and the external robots can avoid themselves without hurting the passenger robot, then the passenger robot should take priority and the robocar should optimize its safety.
The problem with this whole argument is that is supposes the self-driving car has allowed itself to get into the situation in the first place.
For example, the cliff / people scenario suggests that the car was unable to see the problem until it was too late.
Obviously, the answer it to make sure that the sensor range makes these situations almost impossible. There will still be an inevitable situation of some sort, but I suggest it won't be quite the simplistic 'moral' problem that is constantly raised.
From the article:
>Of course, cars will very rarely be in a situation where it[sic] there are only two courses of action, and the car can compute, with 100% certainty, that either decision will lead to death. But with enough driverless cars on the road, it’s far from implausible that software will someday have to make such a choice between causing harm to a pedestrian or passenger. Any safe driverless car should be able to recognize and balance these risks.
An interesting example would be a fast-developing natural disaster, such as an earthquake, that causes falling debris [1] and road damage [2] and as a result quickly and drastically changes available options for many different vehicles and pedestrians in a short amount of time.
4 comments
[ 3.2 ms ] story [ 22.1 ms ] threadNow if the driverless car is empty, then of course, it should gladly destroy itself to avoid hurting human pedestrians, and even any animal or robot.
If the driverless car transports a human or animal passenger, then it should avoid hurting them, even if that means hurting external humans, animals or robots.
But if the driverless car transports a robot, then it should avoid hurting it, even it that means hurting an external robot. But it may hurt its passenger robot to avoid hurting an external human or animal.
Now, assuming some mean of communication between the driverless car and the external robot, and the passenger robot, the driverless car could, if it was absolutely impossible to find any combined action of the external robot and the driverless car avoiding to hurt both of them, then a solution that would hurt one or the other might be found, depending on the criterial of "importance" of each robots. For example, if one or both of them contain information or transports materials critical to save the life of humans or animals in an immediate situation (the reason why one, the other or both robots are moving), then the choice should be to keep moving the one with the best impact.
For example, the robotcar drives a robot transporting the antitoxin that will neutralize a biobomb that could kill 100,000 humans in ten minutes, while the external robot transports the bronchodilator that will save an asthmatic child, and there's no way the robocar and the external robots can avoid themselves without hurting the passenger robot, then the passenger robot should take priority and the robocar should optimize its safety.
For example, the cliff / people scenario suggests that the car was unable to see the problem until it was too late.
Obviously, the answer it to make sure that the sensor range makes these situations almost impossible. There will still be an inevitable situation of some sort, but I suggest it won't be quite the simplistic 'moral' problem that is constantly raised.
An interesting example would be a fast-developing natural disaster, such as an earthquake, that causes falling debris [1] and road damage [2] and as a result quickly and drastically changes available options for many different vehicles and pedestrians in a short amount of time.
[1] https://duckduckgo.com/?q=earthquake+falling+debris&t=ffsb&i... [2] https://duckduckgo.com/?q=earthquake+road+damage&t=ffsb&iax=...