[website] MIT Moral Machine

[Garacesh]

Moral Machine

 

Welcome the Moral MAchine! A platform for gathering human perspective on moral decisions made by machine intelligence, such as self-driving cars.

 

We show you moral dilemmas, where a driverles car must choose the lesser of two evils, such as killing two passengers or five pedestrians. As an outside observer, you judge which outcome you you think is more acceptable. You can then see how your responses compare with those of other people.

 

Bloody interesting. I'm having to have a long, hard think about these questions. I'd hate to be the guy programming this kind of software.

 

Edit:

So apparently I have a slight bias towards upholding the law and saving women over men, a reasonable bias to saving the young over the elderly, a humongous bias to saving human lives over animal lives.

Edited October 14, 2016 by Garacesh

[6Foot2]

...must choose the lesser of two evils...

 

 

[ICTDirect_Dave]

That was interesting. Here's my first pass: Moral Machine

 

edit: A large part of me however was internally screaming that the damn brakes should've been checked!!

Edited October 14, 2016 by ICTDirect_Dave

[Garacesh]

It would have been nice if there was a "I consider either of these outcomes as equal merit.", for example: Should the car swerve and kill an elderly woman, or continue straight on and kill an elderly man.

I can't definitively say one outcome is better than the other as with the information I have, the two choices are equal. I only value gender when the subject is young enough to breed (as women are the primary progenitors of our species)

 

Retook the test in order to get a results page, ended up apparently favouring laws more than previously.

Moral Machine

[gshaw]

Really interesting site that, thought provoking stuff - although the concept of 3 cats and 2 dogs driving a car did make me chuckle

 

Seems the bulk of my decisions were based around the concept that the car should be protecting its owner at the cost of others, along with those not following the rules of the road are the ones I chose to be run over when it was a straight choice of who to run over. Also wonder how the car could ever decide on the social value of a person at any given point in time but then again with pervasive social media and Big Data that's maybe not as far fetched as first thought.

 

Does bring out a very interesting point that if self-driving cars go mainstream how can you put your faith in the appliance you've paid for putting your interests first? Likewise can you be safe in the knowledge of a system making decisions for you without knowing the profile of the algorithm that's been installed? I guess many consumers won't care as long as they get the novelty of being driven around and the marketing is good enough.

[Koldov]

how can you put your faith in the appliance you've paid for putting your interests first?

 

Very true, I should imagine the first port of call would be a hack to 'Save My Life Over Anything Else!'...