Do you feel safer with self-driving cars on the road?

[BWV]

The best way to answer this question is answering this one:

While driving a car, would you decide to kill yourself if you are forced to chose between, say, driving off a cliff or into a crowd of people?

The law does not expect you to sacrifice yourself to save others, but it might require your car to

http://science.sciencemag.org/content/352/6293/1573

 

[Stavros Kiri]

The program called "Thelma & Louise" is especially bad that way.

Here is another question that might question the option for manual override, that I spoke in favour earlier:
Will a human be allowed to manual override an autonomous vehicle to "their death"?

[As a first response, I can't see why not. (Suicide generally may be immoral but not illegal ...)
But what about ... to other people's death?]

 

[jack action]

The law does not expect you to sacrifice yourself to save others, but it might require your car to

http://science.sciencemag.org/content/352/6293/1573

I don't think it will ever happen. The machine is there to make the decision for the driver (now only a passenger), thus if a human driver's reaction is deemed acceptable, the same decision made by AI 'working' for the passenger should be acceptable too.

It would be terrible if human lives were just reduced to probabilities and statistics, because then humans just become livestock. And that is why (from the abstract of your link) no ones want to be in a driverless vehicle that has not its passengers as a number one priority.

Imagine putting your child in a school bus. Could you accept the bus driver sacrificing your child because he might save other (more valuable) people? Now replace the school bus and its driver by a driverless bus. The moral dilemma faints quickly.

 

[256bits]

For the lighter side of winter and self driving cars.

would a self driving car been as "fooled" as much as the police in Montreal, by a snow-car
https://www.autoblog.com/2018/01/18/car-snow-sculpture-montreal-viral-police/#slide-7217631

 

[256bits]

I don't think it will ever happen. The machine is there to make the decision for the driver (now only a passenger), thus if a human driver's reaction is deemed acceptable, the same decision made by AI 'working' for the passenger should be acceptable too.

It would be terrible if human lives were just reduced to probabilities and statistics, because then humans just become livestock. And that is why (from the abstract of your link) no ones want to be in a driverless vehicle that has not its passengers as a number one priority.

Imagine putting your child in a school bus. Could you accept the bus driver sacrificing your child because he might save other (more valuable) people? Now replace the school bus and its driver by a driverless bus. The moral dilemma faints quickly.

The decision would be the choice of the driver, and how a decision made in seconds can play out.

No driver can, and neither could a self driving car, do the necessary calculations in the short time allocated.
Otherwise, there should be enough time to avoid and/or stop harming no one.

Then again, how often do these scenarios ever play out anyways.
But owing to the chance that it could, most drivers would probably try to avoid bus shelters, babies in carriages, bicycles, driving off a cliff, wedding parties, ramming into a building, or whatever. Accidents happen so split second that there is just not enough time to second guess maneuvers. In the end, the casualties may be the occupants of the car, or bystanders, some of both, or neither.

At one time I used to think, ah, that's a moral dilemma to sort out, on how to make the program make a moral decision, or the cost of a life, but not so much any more. If it could be done, ie give the program the responsibility of making moral decisions, question then becomes "Whose morals?" It quickly becomes a quagmire.

Best way to sort is out, as is done now with human drivers, is through the legal system, and payouts, if anyone is ever found negligent and/or responsible for the cause of the accident. Light 'black boxes' could become the norm for cars equipped as self driving as a means to provide evidence.

 

[jack action]

No driver can, and neither could a self driving car, do the necessary calculations in the short time allocated.

That is why there is always a "fail-safe" maneuver which - in the case of humans - is to do what most likely will protect itself. It is hardwired, a reflex. If you use AI, you will have to program one of these maneuvers: In the event I can't decide what to do, what do I do?

For my part, when I face a moral dilemma, I can always count on The Simpsons to show me the way. And here is how The Simpsons believe AI should react in extreme cases:

​

Save yourself!

 

[FactChecker]

Another issue with driving in snow is the problem of staying in the correct lane. That is being worked on (see ).

But this approach requires that the lane signature be mapped ahead of time and available to the car. That doesn't seem very realistic to me. I think it would be easier if something was embedded in the pavement that indicated the lane.