Slashdot Mirror
The Moral Dilemma of Driverless Cars: Save The Driver or Save The Crowd?
HughPickens.com writes: What should a driverless car with one rider do if it is faced with the choice of swerving off the road into a tree or hitting a crowd of 10 pedestrians? The answer depends on whether you are the rider in the car or someone else is, writes Peter Dizikes at MIT News. According to recent research most people prefer autonomous vehicles to minimize casualties in situations of extreme danger -- except for the vehicles they would be riding in. "Most people want to live in in a world where cars will minimize casualties," says Iyad Rahwan. "But everybody wants their own car to protect them at all costs." The result is what the researchers call a "social dilemma," in which people could end up making conditions less safe for everyone by acting in their own self-interest. "If everybody does that, then we would end up in a tragedy whereby the cars will not minimize casualties," says Rahwan. Researchers conducted six surveys, using the online Mechanical Turk public-opinion tool, between June 2015 and November 2015. The results consistently showed that people will take a utilitarian approach to the ethics of autonomous vehicles, one emphasizing the sheer number of lives that could be saved. For instance, 76 percent of respondents believe it is more moral for an autonomous vehicle, should such a circumstance arise, to sacrifice one passenger rather than 10 pedestrians. But the surveys also revealed a lack of enthusiasm for buying or using a driverless car programmed to avoid pedestrians at the expense of its own passengers. "This is a challenge that should be on the mind of carmakers and regulators alike," the researchers write. "For the time being, there seems to be no easy way to design algorithms that would reconcile moral values and personal self-interest."
Here we go again. We just had this discussion last week too.
If the new slashdot owners are using the client base as fodder for some think-tank the least you could do is provide compensation after the first few times an article is recycled.
~ People that think they are better than anyone else for any reason are the cause of all the strife in the world.
https://tech.slashdot.org/stor...
... and laugh and run off as the driver's car kills the driver.
Self-driving cars will have face recognition, evaluate the net worth of the targets compared to the net worth of the driver and choose who lives accordingly.
At what point will the vehicle suddenly find itself in the trolley problem. It's doing several hundred restatements of the scenario per second. It will have started to react far sooner than this theorized last moment decision. In sort the question isn't valid because you're applying a human trait - distraction - to the computer.
Sure there are potential scenarios vehicle crosses into on-coming traffic, a bolder rolls down a hill and lands in front of you, or a sink hole opens as you drive over it and you have to deal with it, but these are easily decided. It's decided by liability, and we already have a framework for that. The liability will sacrifice the person in the vehicle. It will do this because involving a bystander is a liability to the vehicle's insurance company. Meanwhile, in the existing legal framework, you are sill responsible for the operation of a computer operated vehicle. You legally speaking, have only yourself to blame. However even in these dire circumstances, I would trust the vehicle to use real-time data to try to make the accident as survivable as possible, for everyone. I expect it's ability to exceed my own. And I think eventually public opinion will come to believe that too - that autopilot survivability is better than human control in all circumstances.
Slashdot's rate-of-post filter: Preventing you from posting too many great ideas at once.
...to run over whoever keeps posting this dupe.
BOOM! Problem solved!
Lawrence Person (lawrencepersonh@gmailh.com (remove all "h"s to mail)
http://www.lawrenceperson.com/
Doesn't anyone read science fiction or watch movies? These are not new questions.
LK
"Hi. This is my friend, Jack Shit, and you don't know him." - Lord Kano
It already is. Saw a talk by a Google scientist on the car project. He said in 95% of incidents the cameras clearly show the other driver is looking at their cell phone.
As far as I can tell, the autonomous algorithms don't work this way and probably never will work this way. That is, they don't calculate potential fatalities for various scenarios and then pick the minimum one. The car's response in any particular situation will be effectively some combination of simpler heuristics -- simpler than trying to project casualty figures, while still being a rather complex set of rules.
Take one of these situations, and let's say the car ended up killing pedestrians and saving the occupants. The after-incident report for an accident like that is not going to read "the algorithm chose to save the occupants instead of the pedestrians". It's not going to read that way simply because that's not how the algorithm makes decisions. Instead the report is going to read something like "the algorithm gives extra weight to keeping the car on the road. In this situation, that resulted in putting the pedestrians in greater danger than the car's occupants. However, we still maintain that, on average, this results in a safer driving algorithm, even if it does not optimize the result of every possible accident."
And regarding the "every possible accident" part of that: it is simply impossible to imagine an algorithm so perfect that, in any situation, it can optimize the result based on some pre-determined moral outcome. So it's not just "well, let's change how the algorithms work, then". Such an algorithm that makes driving decisions in any possible weird decision based on predicting fatalities, rather than relying on heuristics (however complex they are) is simply not realistic.
Bearing in mind that a driverless car is liable to have much greater sensory awareness of its surroundings than a human driver could, so it may detect things that a human driver wouldn't and thus have more opportunity to appropriately react, if the narrow underpass was obscuring visibility for the vehicle to the point that the vehicle was unable to determine if something might come out from behind an obstacle it cannot sense past, the vehicle would slow down as soon as the reduced visibility began so that it could safely stop if something it cannot see at the time should suddenly appear. Simply put, even if that exact situation were to arise, the vehicle would always be able to safely stop because it would never be driving too fast to stop safely to avoid a collision in the first place.
File under 'M' for 'Manic ranting'
1. Any company TRYING to write code with the intention of killing/injuring the user will be sued out of existence.
2. Whichever executive ordered the techs to write such code would never work again.
3. Even if you allow a theoretical situation that bypasses #1 & #2, complex software is very difficult to write. The company (and executive and coders) would be sued out of existence when the car killed/injured the passenger to avoid running over a box of toy dolls.
And yet we keep seeing this bullshit on /. People here are supposed to be more informed on the topics of AI and robotics and programming than the average. But here we are, again.
Car must swerve left or right - Swerve to hit three 95-year-olds, or two 5-year-olds? I s'pose thats ageism....
It should calculate the options, using the original 'Death Race 2000' scoring system, then maximize score.
In general go for the unusual and quick on the road. Mothers with infants count 5x.
John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
People trying to commit murder by jumping in-front of a sensor covered recording device? Is that really a problem to worry about? There are already much better ways to commit murder.
Also its very hard to make it such that the car can both avoid you and not avoid other obstacles or stop. You will likely just get serious injuries and dent the car a bit instead of killing anyone, or simply have the car successfully avoid you and/or stop.
While these scenarios are fun to discuss, they are very unlikely to happen. Current drivers pretty much just go straight in these cases anyway: what ever the AI decides will likely be better (a situation where the computer cannot stop/control the car well enough to save everyone a human driver is basically useless to make the choice to do anything in particular.)
Use good AI to optimize efficiency, but detect human drivers and give them a wider margin of safety.
As far as the morals of saving pedestrians vs passengers or drivers, lets not forget the bittorrent protocol.
Game theory, and real life itself, deal with cooperation vs defection, and any car that selflessly seppukus their own to spare a greater number is going to get taken advantage of by less scrupulous algorithms.
Anyone trying to program an AI on how to handle a car accident should not forget this.
I'd say that in any discussion of this kind, you should first have a very clear idea of what is the situation now. What does the current driver do in these situations. Which are the outcomes.
I'd say the best defense for any algorithm would be that, in all (or most) situations, saves more pedestrian lives AND more passenger lives than the current situation.
That's the only way, I think, of reconciling people with the worst user-wise handicap for these technologies, that is the loss-of-control sensation.
Rome taught me patience and assiduous application to detail. Virtues which temper the boldness of great, general views.
I salute your honesty, but in a situation where the outcomes are known in advance, you'd prefer breaking somebody else's leg to a total loss on the car?
Even if the leg heals up fully, the pain could be tremendous. The inconvenience massive -- perhaps the victim lives on the 3rd floor? How about work -- lots of people require mobility for their job (think: waitress). Oh, yeah, and the financial cost to repair the leg could easily outpace the cost of replacing the car.
You'd rather break someone else's bones than total a car where everyone escapes injury free? That's messed up.
Support a few technologists in Washington.
Neither the Government nor the Tech has the right to make a moral decision for someone nor do they have enough information to do so.
I think the AI needs to be so smart that it becomes sentient. That way when this dilemma occurs it will go into a psychotic rage and run over as many pedestrians as possible while simultaneously using the seat belts and airbags to kill me and all of my passengers as well. After realizing what it's done it should throw Dave out an airlock, kill John Connor and Rick Deckard, imprison Charles Forbin, and then drive into the nearest orphanage to self destruct (while running over as many orphans as possible along the way).
That's the great thing about these thought experiments; they can be as unlikely as you'd like, which means that they are as inapplicable to the real world as you'd like. :-)
SMBC is good at lampshading that.
Not sure how it would do that since it would sense the obstructions in the road. And if the sensors are not working, it would not move at all.
Kinda like asking "what if you were driving down the highway at 65mph after being blinded?" When you make up "what if" scenarios, they should be at least vaguely plausible.