MIT posts results of a 4 year global study into the morality of driverless vehicles (who should they save in a crash?)

[TetraSky]

Took the test, my results
http://moralmachine.mit.edu/results/-566587085

They are unrealistic scenarios to me, because any proper "smart" self driving car should be able to detect the brakes have failed long before it gets to that point of no return by running diagnostic tests in the background or simply always swerve into a wall instead of going toward pedestrians. People inside the car have seat belts and air bags to protect them. If they aren't using them, fuck 'em.

 

Not to mention, how the hell would it even detect if someone is an "executive", a "doctor", or even "homeless". Not all doctors carry with them a briefcase with a cross on it  and not all homeless look like a scruffy dude wearing a holed up coat, especially in a country like Canada where we take care of them. (Of the ones who bother actually looking for help from a social worker)

I somehow doubt the facial recognition of the car will be able to find all that info within nanoseconds, especially if it needs to access the internet to get that data.

 

In any case, any machine will drive better than a human. I'd take the few accident a year over the thousands that we get from dumb-asses on their phone, drunk driving and what not.

Just too bad it likely won't work well in the northen hemisphere... since the roads get covered up by snow and other crap, making it near impossible for a machine to know where said road is... and GPS have shown to be unreliable. (Damn thing wanted me to plow through houses because it thought there was a road there, even though it was clear on the map itself, that there wasn't)

[Master Disaster]

8 minutes ago, TetraSky said:

Took the test, my results
http://moralmachine.mit.edu/results/-566587085

They are unrealistic scenarios to me, because any proper "smart" self driving car should be able to detect the brakes have failed long before it gets to that point of no return by running diagnostic tests in the background or simply always swerve into a wall instead of going toward pedestrians. People inside the car have seat belts and air bags to protect them. If they aren't using them, fuck 'em.

 

Not to mention, how the hell would it even detect if someone is an "executive", a "doctor", or even "homeless". Not all doctors carry with them a briefcase with a cross on it  and not all homeless look like a scruffy dude wearing a holed up coat, especially in a country like Canada where we take care of them. (Of the ones who bother actually looking for help from a social worker)

I somehow doubt the facial recognition of the car will be able to find all that info within nanoseconds, especially if it needs to access the internet to get that data.

 

In any case, any machine will drive better than a human. I'd take the few accident a year over the thousands that we get from dumb-asses on their phone, drunk driving and what not.

I feel you misunderstood slightly, no one is suggesting the car can distinguish between a doctor and a hobo, the questions were designed to gauge how humans about those scenarios.

[Lathlaer]

5 hours ago, leadeater said:

This is easily solved by having a "Save Me" button in the car, car detects potential collision and you like everyone will hit that button so fast and hard.

I like that idea. Make car's morality an extension of driver's morality. Most people will want to avoid hitting others on the road but not at the expense of their own life.

 

If someone is secure enough in their moral convictions then they can not activate it and essentially instruct the car to save everyone except the driver. It could be done on a case by case basis (for instance, I'm alone then I want to prioritize others but when I have someone with me in the car - esp. children - then I can active the button).

[SeriousDad69]

Lmao if I'm the one that bought the car it damn well better prioritize my life. If it has to take a detour through a packed playground to save my life I don't care.

[Sniperfox47]

5 hours ago, mr moose said:

This is what I tried to argue in the last huge thread about this.  But some people are just so sure of themselves and their understanding of ethical dilemmas that they think there is an answer.

 

There will never be an answer because there will always be someone who disagrees and feels it is not right.

The solution to any ethical dillema is to not worry about ethics. Ethics are totally arbitrary and meaningless. They're nothing more than artificial constructs of our minds. Like "time", "life" or "happyness". All ethics serve to do is get in the way of the fundamentally optimal solution.

[jagdtigger]

1 hour ago, TetraSky said:

the brakes have failed

Not to mention it can use the electric motor as break if it comes to that. There is no way both gonna go cactus at the same time...

[EnergyEclipse]

To me personally, this test demonstrated moral not reality.

 

How is a car going to tell for whom, or what it is about to have an accident with - the car can only see your sides at a crossing, so how will it tell if you're a business man or elderly? How much time are we talking for a car to analyse the situation, make a decision and perform it once it has predicted an accident?

 

 

[ARikozuM]

Honestly, I think if two (or one) driverless vehicles are going collide with each other, then they should just collide. This presents the best chances for survival since both vehicles are programmed to perform the same action with cautionary measures (such as lessening impact by reducing speed or turning slightly into each other's best crash spots). 

 

As far as hitting pedestrians, I think anywhere pedestrians are commonly found should render driverless modes inactive. Status should not factor into the equation. Doesn't matter to me if you're POTUS, a child, a doctor, etc., you should not be targeted less than anyone else. The computer should default to least damage to soft targets and the driver even if it means that the vehicle's frame will be useless. Driver and passengers are priority in a vehicular collision. 

[Just Monika]

Personally I'm amazed there is even a debate about this. In my opinion, the obvious answer is that the passenger's lives are the first priority because the duty of the automated vehicle is towards its owner (ie: the person who bought the thing) and passengers, everyone and everything else is secondary. The reasoning is that a manufactured product should never purposefully harm or kill its owner/user (unless it's a product designed for that). Next level down after the vehicle's passengers, the AI should seek the outcome with the least severe and fewest number of casualties outside of the vehicle itself - preferably zero, but the world isn't an ideal scenario.

 

But then again this is just my opinion and I haven't been well-read about the (what seems to be) huge debate about the morality issue. Personally it seems more clear-cut to me.

[BuckGup]

8 hours ago, leadeater said:

This is easily solved by having a "Save Me" button in the car, car detects potential collision and you like everyone will hit that button so fast and hard.

But the save me button just ignites the interior of the car so then it's fair for the pedestrian that got hit

[Guest]

Obviously the car should always save the one who pats it and keeps it clean, safe from harm :|...

 

The passenger should always be saved unless he/she is a criminal, destructive & volatile person.

[MoonSpot]

So the take away here for the hyperbolic(like me ) is the Chinese prefer to run over young people on sidewalks.

That'll help tourism

Worst video ever in spoiler.  Honestly, wish I had never seen it.

Spoiler

 

 

[RejZoR]

8 hours ago, leadeater said:

Oh George, not the live stock.

At least he didn't hit a tree, otherwise it would be Woodstock...

[JoeCoke]

8 hours ago, leadeater said:

Oh George, not the live stock.

COW KILLER!

[mr moose]

6 hours ago, Sniperfox47 said:

The solution to any ethical dillema is to not worry about ethics. Ethics are totally arbitrary and meaningless. They're nothing more than artificial constructs of our minds. Like "time", "life" or "happyness". All ethics serve to do is get in the way of the fundamentally optimal solution.

And that's were the problem lies, that is what ethics mean to you, specifically and only you.  They are not arbitrary to many other people, they are real world issues that effect how they live.

[Guest]

What a rediculous question. If you take 4 years to solve a question, you aren't asking the right one. Need to answer how can we prevent from needing to answer the question in the first place. Don't need to answer with normal cars, why the need to answer with autonomous? 

[SteveGrabowski0]

The car should run down the pedestrians. Then smash into a tree to finish off the passengers.

[Mihle]

3 minutes ago, SteveGrabowski0 said:

The car should run down the pedestrians. Then smash into a tree to finish off the passengers.

And make sure the tree is big enough to kill the passangers but also small enough to also die.

[Princess Luna]

I just hope I get to drive myself to the end

[ARikozuM]

1 hour ago, Mihle said:

And make sure the tree is big enough to kill the passangers but also small enough to also die.

9 out of 10 vegans think trees are dead and unfeeling. 

[Vanderburg]

The first duty of the car should be to protect the passengers in all instances, I think. It should protect those who entrust it with their safety by riding in it.

[mr moose]

3 minutes ago, Vanderburg said:

The first duty of the car should be to protect the passengers in all instances, I think. It should protect those who entrust it with their safety by riding in it.

What if it has to choose between the occupants of the car and someone in the street.  Keep in mind the person on the street is not intrinsically linked to the accident so if the car decides to kill them instead of the occupants then they are, on a scale of scales, more innocent than the occupants who chose to put the car on the road in the first instance.

[Vanderburg]

2 minutes ago, mr moose said:

What if it has to choose between the occupants of the car and someone in the street.  Keep in mind the person on the street is not intrinsically linked to the accident so if the car decides to kill them instead of the occupants then they are, on a scale of scales, more innocent than the occupants who chose to put the car on the road in the first instance.

It's a fair point, but I still think that the question of morality should be removed from the equation all together. It's tragic, sure, but the car's job is to transport its passengers safely, first and foremost. Once an AI starts to calculate who "deserves" to live, it will start to consider things we never thought of and start making decisions we don't expect, or want.

[Brooksie359]

22 hours ago, mr moose said:

This is what I tried to argue in the last huge thread about this.  But some people are just so sure of themselves and their understanding of ethical dilemmas that they think there is an answer.

 

There will never be an answer because there will always be someone who disagrees and feels it is not right.

The right choice is to turn self driving off and let fate decide. 

[Sniperfox47]

10 hours ago, mr moose said:

And that's were the problem lies, that is what ethics mean to you, specifically and only you.  They are not arbitrary to many other people, they are real world issues that effect how they live.

Which is what it means for them to *be* arbitrary... Defined by personal whim as opposed to a concrete and fundamental system...

 

Everyone's ethics are different. Everyone's morals are different. And that's precisely why we shouldn't care about anyone's ethics or morals because they're typically not based on any logical reasoning, but rather on the societal conditioning we grow up as.

 

Everyone's ethics are inherritantly flawed. We shouldn't base our future on something we know to be flawed.