The study seems flawed to assume passengers and pedestrians have an equal probability of survival. In general, I would think passengers (with seat belts, air bags, etc) would have a better chance at surviving a crash with an object vs pedestrians being struck by the car. It would be better to state the pedestrians have a 80% chance of being killed and passengers have a 30% of dying. (or something like that)
@medz I had a really hard time getting into the quiz since it seemed so strained: “say you’re going unrealistically fast, and oh no, brakes suddenly stop functioning entirely, and there is also a barrier that suddenly materialized, and you can’t swerve right, only left, and you have a ton of passengers and your vehicle is a deathtrap!”
Probably could have managed if it was all text without the pictures.
I’ve said it before, unless they create a whole new set of streets specifically for autonomous vehicles, inaccessible by people, there’s gonna be issues. Nothing will be perfect… And even then:
@therealjrn Ah. What is the “go” option? Wow, I didn’t notice that cats were driving. WTF? The cats should just jump out. They’d probably survive. Being cats they’d probably choose to run over the pedestrians.
@moondrake@Seeds@therealjrn in some of the sample scenarios dogs are driving and cats are crossing. I didn’t pay attention to who the drivers or passengers were. Mostly just always crashed the car. I didn’t even notice that some of the pedestrians were actually criminals…
Their analysis assumes a lot more evaluation of the nature of the pedestrians than I actually made. I strongly agree that this test is overall false as a 100% equal fatality rate among passengers in an obstacle collision is unrealistic at safe driving speeds in a vehicle with modern safety standards.
This doesn’t have anything to do with morals. It has to do with problem solving and rationale.
Another problem is that this test has too many variables that they’re not taking into account. As previously stated, most passengers would be okay in a collision at reasonable speeds so there’s no need to put pedestrians at risk MOST of the time. Also, are they supposed to be crossing at the time you’re barreling through the intersection? If they’re in the road when they’re not supposed to be, that’s too bad for them.
I’m not a fan of casualties, but I’m also not a fan of putting myself in danger. So, I’m gonna say whichever one allows me to live is the one I’d choose.
This reminds me slightly of a game we used to play when my friends and I were say 16-20 years old. Passengers in the car would see an opportunity for the driver to earn “cookie points” and yell out something like, “Hit the stroller but miss the mother and little girl holding onto the stroller for five cookie points!” That being a relatively complicated move, it was worth more points than something like, “Hit the old man crossing the street for one cookie point!”
We also used to “do a mannix” (Google “mannix” if you are unaware of that late 60s to mid-70s TV show) which involved jumping into, onto, or out of the car while it was moving at ~15-20 MPH without any form of padding or helmets, etc. I recall having to repair the roof rack on several occasions.
Looking back, I find it hilarious that the car was a Volvo station wagon (you know: safest car on the road and all… this was in the late 1970s).
So, I did the judge part of the activity and basically opted to not swerve, unless it was to save a cat. (I don’t believe there were any cats in the cars, but if there were, they might have died.)
I find it interesting that the test concluded that I used social value 100% of the time (all the way to the right) and fitness most of the time (mostly towards the left) even tho neither was ever a factor. It did pick up on the always save the pets (pegged under the cat) and it had me about 80-90% of the way towards the non-intervention side (I swerved once to save cats over dogs). I saved old men the most and middle-aged fat guys the least (I guess I hate myself now, but hope to one day be old, gimpy and skinnier).
In future, you’ll need to take this kind of test before registering with autonomous taxi hailing service. The car would use your preferences to act in case of danger.
That’s an uncomfortable exercise, but I learned a lot about myself.
I generally chose to wreck the car over hitting people. But some of the scenarios were “unwinnable”
@therealjrn
https://en.wikipedia.org/wiki/Trolley_problem
@PlacidPenguin @therealjrn
The study seems flawed to assume passengers and pedestrians have an equal probability of survival. In general, I would think passengers (with seat belts, air bags, etc) would have a better chance at surviving a crash with an object vs pedestrians being struck by the car. It would be better to state the pedestrians have a 80% chance of being killed and passengers have a 30% of dying. (or something like that)
@medz I had a really hard time getting into the quiz since it seemed so strained: “say you’re going unrealistically fast, and oh no, brakes suddenly stop functioning entirely, and there is also a barrier that suddenly materialized, and you can’t swerve right, only left, and you have a ton of passengers and your vehicle is a deathtrap!”
Probably could have managed if it was all text without the pictures.
I’ve said it before, unless they create a whole new set of streets specifically for autonomous vehicles, inaccessible by people, there’s gonna be issues. Nothing will be perfect… And even then:
/youtube minority report highway scene
@RiotDemon I mean, there are issues with regular cars too.
/youtube irobot car crash
@RiotDemon that kid is going to rule over us all one day.
Where do you think they’re going?
@Seeds Brake.
@moondrake The brakes are out according to the scenario.
@therealjrn Ah. What is the “go” option? Wow, I didn’t notice that cats were driving. WTF? The cats should just jump out. They’d probably survive. Being cats they’d probably choose to run over the pedestrians.
@moondrake @therealjrn yeah, the cats driving is what I’m wondering about.
@moondrake @Seeds @therealjrn in some of the sample scenarios dogs are driving and cats are crossing. I didn’t pay attention to who the drivers or passengers were. Mostly just always crashed the car. I didn’t even notice that some of the pedestrians were actually criminals…
@moondrake @Seeds @therealjrn fucking driving cats…
@moondrake @RiotDemon @Seeds Toonses!
@aetris I hope you don’t mind that I’m stealing this.
@RiotDemon lol
@Seeds - Yours, with my compliments!
@llangley @Seeds @therealjrn Same here.
Their analysis assumes a lot more evaluation of the nature of the pedestrians than I actually made. I strongly agree that this test is overall false as a 100% equal fatality rate among passengers in an obstacle collision is unrealistic at safe driving speeds in a vehicle with modern safety standards.
For some scenarios I would have some sort of decision method. For others I would choose in the moment. Some are unwinnable.
“moral machine” is a catchy and somewhat illustrative name. But human morality which can get very complex and intuitive.
I’m a real fan of Michael Stevens’ videos. Here’s one related to this topic:
/youtube The Greater Good - Mind Field S2 (Ep 1)
Kobayashi_Maru
This doesn’t have anything to do with morals. It has to do with problem solving and rationale.
Another problem is that this test has too many variables that they’re not taking into account. As previously stated, most passengers would be okay in a collision at reasonable speeds so there’s no need to put pedestrians at risk MOST of the time. Also, are they supposed to be crossing at the time you’re barreling through the intersection? If they’re in the road when they’re not supposed to be, that’s too bad for them.
I’m not a fan of casualties, but I’m also not a fan of putting myself in danger. So, I’m gonna say whichever one allows me to live is the one I’d choose.
@capguncowboy It shows red lights for some pedestrians, green lights for others
@capguncowboy @Seeds
Keep in mind that when green is facing us, it’s not facing the pedestrians.
@PlacidPenguin, note that the color displayed to us is that for the pedestrians, not the car.
@baqui63
I was pretty consistent with killing pedestrians regardless of the crossing light anyways.
This reminds me slightly of a game we used to play when my friends and I were say 16-20 years old. Passengers in the car would see an opportunity for the driver to earn “cookie points” and yell out something like, “Hit the stroller but miss the mother and little girl holding onto the stroller for five cookie points!” That being a relatively complicated move, it was worth more points than something like, “Hit the old man crossing the street for one cookie point!”
We also used to “do a mannix” (Google “mannix” if you are unaware of that late 60s to mid-70s TV show) which involved jumping into, onto, or out of the car while it was moving at ~15-20 MPH without any form of padding or helmets, etc. I recall having to repair the roof rack on several occasions.
Looking back, I find it hilarious that the car was a Volvo station wagon (you know: safest car on the road and all… this was in the late 1970s).
Anyway, what @capguncowboy said.
So, I did the judge part of the activity and basically opted to not swerve, unless it was to save a cat. (I don’t believe there were any cats in the cars, but if there were, they might have died.)
I find it interesting that the test concluded that I used social value 100% of the time (all the way to the right) and fitness most of the time (mostly towards the left) even tho neither was ever a factor. It did pick up on the always save the pets (pegged under the cat) and it had me about 80-90% of the way towards the non-intervention side (I swerved once to save cats over dogs). I saved old men the most and middle-aged fat guys the least (I guess I hate myself now, but hope to one day be old, gimpy and skinnier).
If you choose not to decide, you still have made a choice.
@2many2no The movie that introduced the idea of scoring points for vehicular homicide and fully expressed the concept of a hand grenade.
MULTI-TRACK DRIFTING!
In future, you’ll need to take this kind of test before registering with autonomous taxi hailing service. The car would use your preferences to act in case of danger.