Ethical dilemma on four wheels: How to decide when your self-driving car should kill you

Los Angeles Times ^ | 06/23/2016 | Karen Kaplan

[HiTech RedNeck]

That speaks for itself.

[HiTech RedNeck]

And with a split second to evaluate what you have, how do you know this is “premeditated suicidal idiot” or “some poor soul in a panic?”

I would take the risk of evasive maneuvers if there was the smallest question.

I believe you are the one lacking the common sense.

[BitWielder1]

We need some version of laws of robotics.
I think "Protect YOUR Passengers." should be on top of the list.
After all, human drivers are not required to sacrifice themselves.

[Scrambler Bob]

Isn’t this like one of those cat in a box mind experiments?

Answer will depend on if you have dark tinted windows, and if you open a car door.

[Covenantor]

Real life scenario:

Los Angeles, intersection of Florence and Normandie, April 29, 1992.

Tractor trailer rig approached the intersection where youts were rioting.

What does the google car, Elon Musk Tesla, Mercedes-Benz driverless car do when faced with a stop light as actually happened?

Similar situations occurred in every major city and town across the country that day as well as in years before and after, so it's not a 3arare thing at all.

Build huge suburban research campuses isolated from real world situations concerns and your narrow focused, barely socialized academics and wizards will fuck up royally to feed their visions.

[DesertRhino]

“The reluctance to sacrifice is foreign to true Christianity, for whatever that observation is worth.”

Goodness, that’s kinda a hard blow. I don’t think i’ll be able to eat again until noon or so.
As for it being “unchristian” in your opinion, that kind of sinful thing is -precisely- the kinda thing I am likely to do that Jesus has already redeemed me for.

And say what you will, i’ll never get in a car that will decide I need to die to save someone else. Period.

[rlmorel]

More likely, you will be preferentially steered through these neighborhoods, because avoiding them would be “profiling”.

Then, if your car breaks down, there will a large flag and strobe light to proclaim your helplessness.

Think of it as a magnet.

[RedStateRocker]

Well, if all cars were self driving the routing could easily be many times more efficient. Consider how much time is wasted at intersections because of sub optimal timing or sensing. Pretty easy for a program to take the data of all of the cars and route everything in the best manner, if I understand correctly that's pretty much how data packets move on the internet so the math is well known.

I think self driving cars are going to replace taxis first. They will be really expensive to start, but the economics of not having to pay drivers will make economic sense in densely populated areas even if a self driving Prius costs $200k as it will be able to function 24/7 except for fueling and maintenance.

They might make a lot of sense in cities where they can be coordinated with each other, lights, bridges, crosswalks exactly like a data network. Next will probably be big rigs, perhaps with a hub outside of the city area where a human driver can board and take the rig the last few miles.

I rather like the idea of some of this functionality for the straight stretches of long trips, slightly slower speed would often be made up for by being able to read or sleep for a bunch of hours on a long interstate drive; 80 through Nevada, for example.

[rlmorel]

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.

2. A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.

3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Laws.

"..But...but...if I avoid the bicyclist in my lane, I will drive my owner over the cliff, and if I do nothing, I hit the human on the bicycle...no matter what I do, a human will be harmed..."

[cuban leaf]

It’s quite simple. The car’s regard is for the folks within the car. Instinctive, if you will.

How often does this come up for human drivers today? Never, or absurdly rarely.

[cuban leaf]

Self defense?

[Boogieman]

“And with a split second to evaluate what you have, how do you know this is “premeditated suicidal idiot” or “some poor soul in a panic?””

If I only have a split second to evaluate, then obviously they ran out in front of my vehicle and didn’t leave me the option of avoiding them. That’s on them, I really don’t need to know what their motivations were.

[Boogieman]

“If this is your attitude, that you aren’t at least going to slam the brakes hard to try to lessen the damage to the kamikaze pedestrian, then that speaks for itself.”

The question doesn’t allow for the possibility of braking to avoid hitting them. You’re imagining different scenarios than the one that is posed in the article and then condemning people for failing to respond to the imaginary scenario in your head. Get a grip.

[IronJack]

There would be no SPEED inequality. If the direction vectors were different, the relative energies could be cumulative. In other words, if Car A was heading west at 30 mph and Car B was heading east on a collision course at exactly the same speed, while there would be no difference in their speeds, the velocity difference would be of great concern to their respective insurance companies.

[BenLurkin]

human drivers are not required to sacrifice themselves

The elites see that as a problem that needs to be solved, apparently.

[Pelham]

Buridan’s Ass. The car will die of hunger and thirst.

[BraveMan]

Will the car be allowed to say it didn't see me?

[I want the USA back]

Obviously, no algorithm can ever substitute for human judgement.
The article raises a valid point.
Here’s what will happen. The government will approve an algorithm, and that will take the force of law. What the algorithm decides will be deemed legal and ethical. Too bad if it decides to kill your mom to save 5 people who you don’t know.
The future.

[detsaoT]

I just dream of the day that an autonomous car can successfully navigate an icy (or heck, even WET) road. It’s one thing to test a car in the perfect weather conditions (i.e., California and Nevada). How well does this crappy software handle the REST of the world?

Stupid software engineers... :)

[BenLurkin]

I heard years ago, don’t know if it’s true, that the original Phalanx air defense systems on our ships would lock up when two missiles were coming at it simultaneously. The computer couldn’t decide which one to shoot first, so it shot neither.

Supposedly the answer was to add a “coin flip” to the computer’s analysis.