Every time I read one of these articles I feel like the author doesn't have a very good grasp of exactly how these work. The car doesn't have to "choose". Nobody at Google is programming these things to calculate the life value of humans, the percent chance of injury to occupants, or the benefits of sacrificing pedestrians.
They're programming them to try and avoid accidents in general. There will be times when the computer fails to do that, but that doesn't mean it decided to kill someone by making a moral choice about human life priority. Maybe in the future this will be a relevant argument but the sort of algorithms mentioned that decide which crash scenario is 'best' simply don't exist.
The best way to deal with these moral issues is to just... not program that behaviour.
They're programming them to try and avoid accidents in general. There will be times when the computer fails to do that, but that doesn't mean it decided to kill someone by making a moral choice about human life priority. Maybe in the future this will be a relevant argument but the sort of algorithms mentioned that decide which crash scenario is 'best' simply don't exist.
The best way to deal with these moral issues is to just... not program that behaviour.