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...
Just ignore the people that vote in the fire hose for more than 3 dupes.
If they voted twice in the first dupe (for each dupe), then voted twice in the second dupe (for each dupe) ,then voted twice in the third dupe (for each dupe)- kill the /. account.
_ _ _ Go for the eyes Boo! GO FOR THE EYES!
I thought there was a way to objectively decide morals: write rules ahead of time. If the car is driving in a perfectly legal manner down its lane in the road, and the 10 people in the road are jaywalking, then the car/driver is in the right of way and should proceed rather than kill its driver. Maybe try to slow down and not hit them so hard, but the car ought not sacrifice its driver for the mistakes of others. You get my point: if you don't want to get run over, then don't jaywalk. Conversely if the vehicle is driving on the sidewalk, and 10 pedestrians are standing on the sidewalk, then the autonomous vehicle ought to swerve into that tree to preserve the pedestrians because you're not supposed to drive on the sidewalk.
... 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/
If you think about it, this would be something manufacturers would consider. Save the driver of a Ferrari 458 or a Mercades S class driver vs the driver of a low budget car like a Honda Accord. They'd never admit to doing this until proven in court. But I guarantee the higher end cars will save the lives of the 1% instead of hurling them off a bridge to save a few pedestrians.
The car should be made to either self destruct in such a situation. Or go ahead and give em vertical takeoff so they can 'jump' over the obstacle.
Oh, and why the frak do these dupes make it on the page? Maybe use some of this vaunted AI to properly vet stories to make sure we haven't just talked about it?
Watson save us!!!!
You fail at physics. There's no way that a single man/woman (fat or otherwise) could stop a train.
If it weren't for deadlines, nothing would be late.
How is that not a moral dilemma of human piloted cars. And seriously, what's the probability that we can debate this day in and day out, and then in the history of autonomous vehicles, it never comes up?
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's going to be a fascinating, if redundant, discussion. The good news is that we will have a long time to discuss it before you start seeing a lot of self-driving passenger cars on our roadways.
Now the real moral dilemma is whether one dollar of public funds should go toward infrastructure for self-driving passenger cars. I mean, if there's money left over once we get back to the point we were at middle of last century, when practically every US city had a robust (and profitable) public transportation system, then we can spend some of that on blue-sky projects for self-driving cars. Then I'd be OK with it.
You are welcome on my lawn.
A driverless car would not drive at a speed that was incommensurate with its own ability to sense things ahead of it which it may otherwise hit. If something up ahead happens to be hidden, whether it is because of nearby buildings or just the topography, the vehicle will be driving slowly enough that it will be able to safely stop *before* it hits something that it cannot yet see. No swerving necessary.
File under 'M' for 'Manic ranting'
For instance, what if that crowd of 10 people were a group of Trump protesters? Then what?
Not that shit again!
Achille Talon
Hop!
I mean it isn't like humanity hasn't been agonizing over these questions since the birth of civilization without coming to satisfactory answers
Welcome to digital morality.
Why is Snark Required?
Nothing immoral about having the car minimize injury to the driver, and fuck everyone else. In most cases this will also minimize injury to those outside the car.
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.
You can stop a train by making a phone call.
If builders built buildings the way programmers wrote programs, then the first woodpecker would destroy civilization.
From a Liability perspective you're safer prioritizing overall minimization of loss of life.
From a Sales perspective, who's going to buy a car that's programmed to purposefully kill you under certain circumstances?
However when the pedestrian is a moose you may want to revise your thinking.
Disturbing pic: http://www.ontario-outfitters....
If builders built buildings the way programmers wrote programs, then the first woodpecker would destroy civilization.
Plenty of obese neckbeards to throw. It's not like anybody will ever miss them.
Irrelevant to the question of whether it actually works. Which it doesn't.
If it weren't for deadlines, nothing would be late.
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.
Put a DIP switch in the car. On position, Save driver at all cost, OFF position, minimize casualties, even if that means sacrificing the drivers. Default it to OFF.
Explain in the manual how to change it.
DO NOT LET THE DEALERSHIP CHANGE IT.
Enjoy safer streets
*** Suerte a todos y Feliz dia!
as long as the tree is not too big
Slashdot, fix the reply notifications... You won't get away with it...
http://www.smbc-comics.com/com...
Imagine you're in an out of control Trolley. You're headed towards three buildings and you control which you slam into. Two buildings contain only one person and one building contains five people. You randomly select a building to slam into. Then one of the other buildings is revealed to contain only one person but you can't switch to that building. Should you switch the tracks to the remaining other building?
Some drink at the fountain of knowledge. Others just gargle.
repeat of a worthless post. get this crap off my screen.
I didn't care before.
Now I want them all dead.
Simple. Google puts a buy back option in the contract for the self driving car. They can buy back your car at any time for the full purchase price. Seems like a swell deal right? They invoke this when you are going to hit a pedestrian to buy back the car. Now it's no longer your car so the choice of who to kill here isn't predicated on your car's loyalty to you. problem solved. Plus, no need for insurance.
Some drink at the fountain of knowledge. Others just gargle.
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.
One person chose to operate several tons of rolling death and metal. Another person just happened to be standing around. The entire responsibility lies with the person who made to choice.
Operate? Choice? Yeah right.
As we consider objects like steering wheels being removed, driving questions around whether or not the rider even needs to be licensed, the future will look more like a rider on a train or bus today. How else are you going to remove the human factor that tends to kill thousands of people every year?
For clarification, reference "autonomous".
This is one of those completely idiotic "thought experiments" that philosophers are so fond of.
I've been driving for 46 years now (I started early, but am still old) and have probably driven well over half a million miles. I have NEVER had to make a decision, "Do I kill myself or swerve and kill ten innocents?"
Almost nearly everyone on /. drives, and I don't even begin to guess the number of miles driven by this audience: has anyone here ever had to make such a decision? Has anyone even heard of such a decision occurring in the real world? For fighter pilots maybe this is a real concern - the Great Santini and the recent Blue Angel who died (as well as others) supposedly stayed with their aircraft to minimize civilian casualties on the ground. But driving a car???
It's about as smart as worrying about what the AI does in the event of a Martian invasion.
Since the second is likely legally equivalent to conspiracy to commit murder I seriously doubt you could take out insurance against it.
Someone cracks, takes the car up to high speed then aims for the sidewalk or swerves into oncoming traffic ?
Car must swerve left or right - Swerve to hit three 95-year-olds, or two 5-year-olds? I s'pose thats ageism....
In roughly a century of driving, humans have learned one strategy: slam on the breaks. The choice is "break, or don't". When the driver is replaced by a bot, the choice is STILL "break, or don't".
I swear, this nonsense about algorithms implementing moral calculus is just a scam to get philosophy professors a few more speaking engagements.
I really doubt this problem would last after all human drivers are replaced.
PS: I don't reply to ACs.
And make it 2 out of 3!
Driverless cars exist. Records of past accidents exist as well. Take those driverless cars and put them into a simulation of those real world accidents of the past and see how they would react. If you find a lot of situations were the car might have saved people by killing the driver, then you can come back and have a discussion, but it's utterly pointless to worry about a hypothetical problem that might never arrive in the real world.
If I get a vote, I'd kind of like a driverless car that doesn't find itself choosing between swerving wildly off the road or hitting a crowd full of people. How does it come up, anyway? I mean, if the car is following the rules, and 10 people spontaneously decide to fling themselves in front of it... fuck it, run 'em down, with a sarcastic little "beep beep" as it drives away.
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'
Mod this up. Very serious issue - if pedestrians know the car will avoid them at the cost of the driver, they'll have no incentive to be cautious. Forget about "look both ways before crossing"; now it's "I cross, you die."
Drive fast enough that pedestrians tunnel through your car.
John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
You should save the people that are actually complying with the law and acting reasonably. Someone crossing the road at a point where visibility is poor and a driverless car can only avoid hitting them by killing its passengers is probably not acting reasonably, and all things being equal, the driverless car should therefore protect its passengers.
So the car roll and kill both the pedestrians and the driver at the same time.
Pedestrians can be legally at fault in car-pedestrian accidents, for example when jaywalking, crossing against a red light, or even by being drunk. In those cases, the pedestrian has no claim against the driver, and the driver of the "several tons of rolling death and metal" can recover damages from the pedestrian.
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.)
The system needs to calculate based on probability of survival. First priority is the driver. No one will buy something that might kill them to save someone else. It should not be up to someone else / something else. After it first calculates best odds of survival for the driver, the next part it factors in is how can it save the most lives without violating the life of the drive, and commit to that action.
Maybe a pedestrian gets run over, maybe it heads in the center of a crowd, hopefully casualties will be minimized in the last second by people jumping out of the way.
At the end of the day, no one is going to sign a blank check to sacrifice their life for someone elses.
What if the group of people were ignoring cars because most of them stopped in time to let them cross whenever they want.
So they try their luck one too many times, and this time the car kills you to save them? I think not.
If it's a "driverless car" then you've left it a bit late to "save the driver".
You're arguing that since it's autonomous, the occupant didn't make the choice to get in? And therefore pedestrians should be killed?
Any pedestrian accepts a level of risk when walking or standing anywhere near what is or will be known as a high-risk zone. (a.k.a. where cars are operating, autonomous or not)
My point was you are a rider, NOT an "operator". You choose to get in a cab today. You do NOT choose which pedestrians it avoids if an accident occurs. That is up to the actual operator of the vehicle (autonomous or not), unless you somehow feel the human riding in the back seat is to blame, all because they needed a ride that day.
In residential streets and downtown areas I always slow down when there is an obstacle I can't see around. Once, a kid actually ran out from behind a parked van I was driving past and it was _only_ because I'd done this that I didn't hit him. I've never hit anyone, nor another car, despite countless close calls over several decades driving, most that were other party's fault and a couple that were mine but I realized my mistake. My brakes have been well-tested.
Anyway I only mentioned that slowing down behavior in reply to mark-t, whose post you must not have read or you wouldn't be asking the question.
The situation was a large cement column on an overpass, from behind which, a group of pedestrians could suddenly emerge, even on the highway. mark-t contended that a self-driving car would simply avoid that by slowing down due to detecting the column but not being able to determine if there might be something behind that could potentially travel into the roadway, and thus it would be essentially impossible for the vehicle to ever get into a situation where it wouldn't be able to safely stop.
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.
If we know the car is programmed to crash into a tree to avoid pedestrian casualties, this can be planned for in the safety design of the car, since it makes the kind of crash more predictable. Further, we can research into how to not get into those situations in the first place. This means looking ahead more when driving (what driving instructors often talk about, what driving students often omit to learn, and what serious police driver training used to drum into people). But being able to compile a comprehensive list of potential accident situations, and a comprehensive list of scenarios leading to them (again research worth doing) and programming measures for each one is not a bad thing. That means that the cars will be predictable in their behaviour in the case of an accident, and that predictability can be harnessed in programming other cars so as not to make things escalate (as happens on motorway pileups all too often: a driverless car will be programmed not to make the silly mistakes too many human drivers do). The best human drivers probably are better than driverless cars, but I imagine the median driver is much worse, and faced with driver safety and crowd safety will often end up achieving neither.
John_Chalisque
The car will be programmed to take whatever action minimizes the manufacturer's liability.
Please stand clear of the doors, por favor mantenganse alejado de las puertas
We already have laws around these things - that dictate what a driver is supposed to do in these conditions and what degree of liability he would have towards passengers or pedestrians. Autonomous cars should do exactly what the local law would have demanded a human driver do.
Unicode killed the ASCII-art *
Define the propensity for the AI to hit a pedestrian by the road type:
1. Limited access roads (highways): If you're on the highway the AI will hit you. It's called limited access no pedestrian for a reason.
2. City roads and side streets: run into a tree. Cars have all sorts of safety features including an air bag. A tree hit won't kill you. Have a dash board camera to record the face of the pedestrian and have them be held responsible for the crash.
3. New slashdot owner walking across a city road: hit him until he is a pile of flesh a la Doom 1.
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.
What else?
Add some randomness into the algorithm just to make it more like real life.
It's hard to say that one decision is always correct, so choose differently from options presented.
If I were to own such a car I would expect it to set my own safety as its number 1 priority. After all he paid for the car and as such expects the car to protect him. I would not buy a car that would do otherwise. If I have to choose between my own life and that of another then I choose my own: I do not want to die. But you could also let people go through an option menu allowing them to adjust such priorities.
I don't see why this is such a conundrum. Right now we presume the driver of a human operated vehicle will in most cases attempt to save the occupants of the vehicle first since the imperative of the driver will be self preservation. I see no reason why this would need to change. All that has changed is that the driver isn't human but it's reasonable to expect the driver of the vehicle (human or not) to attempt to preserve the life of the occupants of the vehicle first because it fundamentally will have more control over that. People outside the vehicle should be responsible for preserving their own lives because they have more control over that. Basically preserve the status quo. The fact that a machine is driving really doesn't necessitate a change. Yes there will always be some corner cases but that's no different than today either.
No matter what we do there will be some casualties just like with almost any new technology while people get adjusted to it. Fatalities in automobiles, aircraft, trains, etc all were substantially higher when they were new technologies than they are today. Over time it will mature and legal frameworks will be built to deal with the accidents that will inevitably occur.
You fail at physics. There's no way that a single man/woman (fat or otherwise) could stop a train.
Really?" You sure about that?
Everybody leaves out responsibility. It makes it easy, if the child is on the road, then hit the child. If the child is in a playground, then kill the passenger. And before someone says children aren't responsible for themselves, no, they largely aren't. It is the parents responsibility in this case. 50 people running onto a motorway should not result in 5 dead passengers as the cars zoom over a cliff, it should result in -insert number- dead people.
I reserve the write to mangle english.
This lack of a decision should always be legal.
What if it's Superman in that one story where he got super-fat?
But what if all of those 5 people are Hitler? Would you go back in time and prevent the invention of fast food to kill 5 Hitlers?
Forget magic. Any technology distinguishable from divine power is insufficiently advanced.
Is there a reason the car has to rely only on its onboard sensors? Since we're putting closed-circuit cameras everywhere, we can as well let nearby vehicles tap into them to see behind corners. As long as the car knows where the camera is relative to its own model of the surroundings, it should be easy to incorporate the image (or preferably preprocessed spatial information) it sends into said model.
In fact autonomous vehicles should definitely link up with each other and surrounding infrastructure. That would allow each vehicle much better situational awareness and help smooth traffick jams both because cars can take alternative routes and because better coordination allows higher throughput.
Forget magic. Any technology distinguishable from divine power is insufficiently advanced.
If I'm paying for the car, it had better be looking out for me.
quiquid id est, timeo puellas et oscula dantes.
More importantly no law should ever say I must sacrifice my own life to preserve others.
I think you'll find the Uniform Code of Military Justice has some exceptions to that but in civilian life you are quite right. I cannot imagine getting into a vehicle where it might be programmed to preferentially kill me instead of someone else depending on the circumstances. If I'm the driver of a vehicle and the tragic choice is my life or yours, unless you are my wife or child you are going to die. Sorry about that but there are only a few people whose life I value more than my own. Chances are yours isn't one of them and I would expect the same in reverse. The fact that a computer is driving the car instead of me does not change that calculus in the slightest.
I've said it before, and I'll say it again for your benefit now. The car is going to do its best to not break the law. Hitting a car that cut you off is not against the law; it's the fault of the driver who cut you off and hit the brakes, that's called a brake job. It's not going to come around a blind turn and hit something because it's going to slow down for the turn if some other vehicle hasn't just gone ahead and reported back that it's clear via V2V. It's not going to leave the lane because that would be a potential violation of the law. It's going to hit the brakes, and diminish the force of the impact. It's not going to choose to swerve around a bus and hit a bunch of nuns, but it's also not going to choose to hit the bus in the first place by driving like an asshole.
If anyone wants to bring up this point in the future, and I strongly suggest that you do not, please say something interesting or novel about the situation.
Peter Dizikes at MIT News is a staggering idiot if he thinks that the majority of people will not buy self-driving cars simply because they might kill you. People also get on airplanes and they can't even get into the cabin and fight with the pilot any more.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Your car is filled with airbags and seatbelts and crumple zones and all sorts designed to protect you during a crash. Pedestrians have none of that (at least for the time being)
All those safety features are independent of the decision of where to aim the vehicle. You never use them intentionally. They come into play when the decision of where to aim the vehicle goes awry. Pedestrians lacking protection is nothing new and does not change the situation in the slightest when considering whether to put computers in charge of the driving.
The CAR should protect you (using those safety features), the AI should do what drivers are supposed to do - cause the least amount of carnage on the road.
It is unreasonable to expect a driver to protect the lives of others in preference to his own. We don't expect that now and having computers drive the car does not change that calculus. There is no way I would get into a car that was purposely programmed to preferentially kill me as an occupant under any circumstances. I'm willing to get on public transportation only with the understanding that the pilot/engineer shares my sense of self preservation. If I thought that there was any chance that the driver a vehicle (human or not) would not protect my life then there is no chance I board that vehicle.
Sorry but if the choice is you or me then you are going to die. Nothing personal. I would expect the same in reverse. The only people whose lives I value more than my own are immediate family and you aren't a member of that group. Like they say before a boxing match, protect yourself at all times.
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.
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?
First of all, I'm having a really hard time imagining even a remotely plausible scenario in which those would be the only two possible options. Has anything even vaguely similar ever actually happened in the real world (that didn't involve the driver driving like an idiot in the first place).
Secondly, it seems to me that even a remotely intelligent AI would be unlikely to get itself in such a quandary in the first place.
Still, if we're going to play the game, I say head for the tree. The pedestrians didn't choose to submit the preservation of their lives to an AI, so even if there's just one (and somehow all other directions are equally deadly), risk the driver instead of the pedestrian. In almost all cases they'll be far more likely to survive.
systemd is Roko's Basilisk.
Wrong!
Kill the driver, every time!
It's an autonomous car, who cares if the computer driving it dies.
Now, the passenger on the other hand...
My karma is in a nose dive
Driverless cars will be able to apply maneuvers to prevent incidents. Eventually these vehicles could go faster than 300mph in the city and avoid obstacles perfectly due to high speed reaction. The scenario postulated though is not possible once the tech advances because the car would be able to navigate through the potential of collision and with advanced anti-grav modules for passengers inside, they won't even spill their coffee or be anywhere near aware of what may have just happened.
Computers can drive exponentially better than flesh.
The dangers of knowledge trigger emotional distress in human beings.
It is easy to argue that a perfect driverless car should act according to strict utilitarian principles, maximizing the number of lives saved. But a perfect driverless car, bug-free and unassailable, is still decades away, if it is even achievable. Imperfect driverless cars are close, but the rules are different. They must be. Until it can be proven that a driverless car is bug-free and utterly immune to outside attack, there must be no code path that allows it to deprioritize the lives of its own occupants. The three reasons for this are simple: bugs, attacks, and buggy attacks.
The issues with bugs and attacks are clear: bugs can cause random deaths, and attacks designed to kill the passenger create a new tool for those who would murder. But buggy attacks -that is to say, attacks that are not designed to harm anyone, but do so anyway because faulty attack code- may be the biggest threat of all. More than one piece of malware, particularly among the early viruses and worms has proved far more destructive than its creators ever intended, all due to bugs, not in the code of the system being attacked, but in the attack code itself. Even if the code in a driverless car's system can be guaranteed bug-free, we cannot assume this of attack code, which is what makes immunity to attack so important.
We are not at a state where we can guarantee such security. Until we are, we must not allow driverless cars to deprioritize their own occupants' safety, even in cases where doing so holds great philosophical appeal. Doing so would almost certainly take far more lives than it would save.
Unless the people are wearing outfits, it's doubtful that a human would know that one person is a Nobel Winning Physicist and the other is a NeoNazi. (twist: a person could be both)
Any true "moral dilemma" would have to include information about all the parties involved.
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.
Should a Nobel Winning Physicist be sacrificed in order to save a car load of Gang Bangers on their way to a Drive By Shooting?
What if that Hypothetical Crowd was a bunch on NeoNazis?
Morality is a Human thing and can't be programmed.
And there you touch upon some of the messy moral morass. The US, in increasingly reactionary fashion, demands fault and punishment, and this will not go away simply because it is a computer making the decision.
This will be the hardest problem to solve, even if the AV fans deny it is a problem at all. No doubt at all that accidents will be reduced, but mechanical components fail, road conditions can become treacherous, and not all animals or children have autonomous control.
And since the human, which is now just a passenger in a vehicle he has no control over, can hardly be held responsible for any accidents, not only do we have a dillemma that does not fit in with our social mores, but will need a remarkable shift in attitudes, one that does not fit with our reactionary nature If the hottie on the weather channel is telling people they should stay indoors, should the car refuse to start ala "I'm sorry Dave, I'm afraid I can't do that!"? The list just grows longer the more we think about it.
And anyone belittling or dismissing the problem simply isn't getting it.
The shepherds did so well protecting the flock that the sheep no longer believed that wolves existed.
Cases like this are why we still have human drivers and should continue to let them have the final override.
Twitter supports and protects racists - by smearing their critics with the "Hate Speech" label.
Lets say you are driving your old, 2001 model car. The brakes go out, would you aim for a crowd of soft people? No of course not. You do your best to miss everything, eventually you hit something, most likely a wall or other barrier - after you intentionally AVOID the people.
Because that is exactly what the idiots posing this question are talking about. The car will not be programmed to ram into people to slow down, no matter what the circumstances.
excitingthingstodo.blogspot.com
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).
One of the many posts on the subject of RTBL, which did almost as you suggested.
Twitter supports and protects racists - by smearing their critics with the "Hate Speech" label.
The problems get very difficult when the cars' choice of actions are determined by the interests of the car manufacturer (and possibly with the insurance company - if those are still different entities) rather than by the occupants.
For example, it seems likely that when there are a reasonable percentage of autonomous vehicles out there, they will be able to communicate with each other - that's a handy thing for negotiating who goes first at intersections and for crash-avoidance.
So now there are a whole raft of other moral dilemmas at levels far below "How many people die?":
* Should the car that knows that its owner is late for work go first at intersections?
* Can you pay more for a car that gets preferential treatment at intersections?
* On a freeway, can cars choose to slow down to save gas or speed up to get there faster? How does this work when cars are "drafting" to save gas? In a "road-train", who gets to decide the speed of the train?
* Are cars allowed to lie to other cars?
Then there are issues about low level accidents - where no humans are harmed:
* In the event of a choice between a fender-bender with car A and car B - can your car figure out which one will cost the least to repair? Will this result in more crashes with cheaper cars?
* Will insurance companies insist on cars choosing outcomes that minimize their liability?
And when humans are harmed:
* From an insurance perspective - it can be cheaper to have the occupant of a vehicle die than having lifetime health issues caused by the accident. If volkeswagen will fake emissions figures and indirectly cause a bunch of people to die as a result - who's to say that some car company/insurer won't bias the AI's parameters to save them money?
It's naive to assume that accidents will cease altogether when AI drivers take control - cars are complicated machines - and parts break unexpectedly all the time. If your brake line suddenly ruptures, then the AI's expected stopping distance is shot to hell - and someone can still die. The AI will need to make life and death decisions as well as broken-rib-versus-crushed-ankle and my-insurance-pays-versus-his-insurance-pays choices. Only now we add safety-reputation-of-my-manufacturer versus safety-reputation-of-competitor decisions too.
This is going to get difficult! Lawyer up!
www.sjbaker.org
Generally speaking, highways don't tend to have lots of places where there is reduced visibility, or else it would be generally unsafe to travel at high speed there in the first place. Where such reduced visibility occurs on a highway, perhaps as you approach a viaduct from a lower level or some such thing, you'd probably notice roadside signage to that effect that recommends that drivers reduce speed anyways, If there are spots up ahead on the road that the car cannot see past, a driverless car will simply not drive fast enough that it will be unable to stop before such places, which may ideally be how a human should be driving too. Consider that humans can't generally make reasoned choices about too many things simultaneously, which would create hard limits on how fast a person might be able to safely drive in such an area, but those limits would not be the same for a computer, which can perform many calculations about proximity and safety multiple orders of magnitude more quickly than any human ever could. The fact that people often may not slow down in such circumstances is a consequence of the fact that human beings often make assumptions that are founded on irrational feelings or intuition instead of objective and logical reasoning. Any annoyance that might be felt at such slowdowns, which are already liable to be infrequent on a highway anyways because reduced visibility isn't the typical state of affairs on such roadways, would only be a consequence of being perhaps accustomed to unsafe driving practices, and not because the car is actually driving poorly.
File under 'M' for 'Manic ranting'
Because the distance you can see determines how fast you can drive safely? If this is news to you, you really shouldn't be on the road.
Decrease speed as significantly as possible to minimize injury/death to the pedestrians and take whatever standard evasive action possible before the collision happens. If hitting the pedestrian is unavoidable (like, because they ran out in front of the car) then tough beans. In any case automated systems have vastly better (nearly instant) reaction time compared to a human driver so the stopping distance is really only limited by the mechanical capabilities of the brakes. In this super-contrived example with a human driver you would probably have 5 dead and 3 critically injured from a human driver vs. say 2 dead and 2 critically injured from a self-driving car just due to differences in impact speed from reaction time decreasing. It's a non-issue. Most people don't have the moral awareness to drive themselves into a tree to preserve "more lives" when faced with a snap judgement call. We shouldn't expect it of our AI's either.
Plenty of obese neckbeards to throw. It's not like anybody will ever miss them.
Irrelevant to the question of whether it actually works. Which it doesn't.
Yeah I'm still not convinced. This really should have been the topic of a Mythbusters episode.
I need to be convinced with some explosions, preferably with lots of fat person body-parts flying around.
In the free world the media isn't government run; the government is media run.
William Shockley
William Shockley
That's often the case, but not always.
You're driving on a highway with no shoulder and one lane for each direction of traffic. The is an oncoming semi and you are right at the start of a narrow underpass. Now a group of pedestrians walks out from behind a support beam and in front of your car. Hit the support beam or hit the pedestrians or hit the semi. Lose, lose, lose.
Were the 3 "lose's" for the 3 pedestrians? Because they had no business being there and therefore they can die.
In the free world the media isn't government run; the government is media run.
I will NEVER be found in a 'self-driving car', unless it has a full set of manual controls and is driver-centric, with any so-called 'autonomous driving' mode as merely a sophisticated Cruise Control that I can completely turn OFF. One way or another. You jackasses who think we'll have boxes on wheels with no controls for a human operator are off your meds and need to be restrained, IT WILL NEVER HAPPEN. EVER. Get over it. At the very least I'll return to the days of my 20's when all I had was a motorcycle; I'd rather ride in the driving rain than EVER get into some deathtrap on wheels that I can't control with my own two hand and own two feet.
Are YOU using the TOOL, or is the TOOL using YOU? Think about it!
Which is why I'm not riding in a driverless car at this stage.
To butcher a quote from the Matrix: "when you're ready, you won't have to swerve"
If I have been able to see further than others, it is because I bought a pair of binoculars.
In that situation the only acceptable solution is to execute a bat turn and recalculate a path through the saner part of the city.
If I have been able to see further than others, it is because I bought a pair of binoculars.
Yeah, no. People stumble suddenly into the road. Pets dash across, as do children. Wild animals can arrive in-the-way faster than you'd believe possible if you haven't actually experienced it. Even really big ones.
The car won't be inattentive, and so it's a fair assumption that it should be able to do better than we would, but it isn't going to be able to avoid everything. At times, choices will have to be made.
Frankly, my position is "save the occupants of the car", because in the case of humans, we can tell then to stay the fuck out of the road and maintain positive control of their damned crotch-blossoms, and they should do so. If not, it's on them, and in the case of animals, much as I love them and feel that we should be their stewards and not their butchers and murderers, I still lean (barely) towards "save the occupants of the car first."
Of course, I also think we ought to make any roadway that carries traffic above 25 mph impossible for bicyclists, pedestrians, pets and wild animals to access, but Those In Power have decided that the number of lives lost to such open access is, in the final analysis, acceptable, as compared to the costs of making it happen. So I'm back to "save the occupants of the car first."
Hopefully we'll have our flying cars soon and creatures can walk the land without fear of such hurtling dangers. There's this fabulous thing, for instance.
I've fallen off your lawn, and I can't get up.
Sweet Jesus, this again?
This thing isn't being built by philosophers, it's being built by engineers. Any oh shit scenario results in the car decreasing speed and coming to a stop. No swerving. No leaping off bridges. No moral dilemmas between measuring the value of a bus full of nuns over the value of it's passengers. The policy is straight forward: Something goes wrong, hit the brakes. This is what they teach everyone outside of California. And because there's also some pedantic asshole who thinks he knows how to drive better than everyone else, it's not going to slam on the brakes, it's going to try and stop in a controlled manner. Stop, just stop. Unjerk the knee. You don't have to comment on it. If someone or something is in your fully legal right of way, then they'd be fucked just as much as if a human were driving. The car will hopefully see it and stop. Just like a real boy.
The one exception I've found is rail-road tracks. Oh shit still means slow down and stop. But if you find yourself on tracks, creep off. If there's an accident and you can't get off the tracks, hell, that warrants a call to 911.
Who are these people that keep trying to desperately insert some moral/philosophical/AI dilemma into self-driving cars?
A bunch of innocent flower children in the road? No a more likely question is what should happen with a violent mob of people? The real problem is when 20 to 100 people decide they want to fork something up and you're there. So that's why I don't like the idea of a car that disables itself, because not going to determine if these are malicious people or not. Years ago I had a group of 100+ drunk college party kids stop my car because I was on "their road" and decide to rock it and smash it up. I let them have their fun and I didn't run down the adults in front of my car. But what would have happened if they did that to a horse? And where I live now there is sometimes gun fire and gang fights in Holyoke Massachusetts on streets I drive to work. I have my doubts about the kind intentions of these gang members and mobs. The real question should be what does it do when a gang want's to stop your car? Will there be an override mode? And do you think any politicians car will stop for a group who has bad intentions for them? Heck, the politicians car will probably aim for the softest looking people. So we'll need politician mode too right?
Customer choice. Drive selects preferences. We get the benefit of more safety than a driver and people will but the technology. Some may even select of sacrifice themselves at a specific level. 1 to 1, Save my ass. 10 to 1 save them.
Why is it so hard to only have politicians for a few years, then have them go away?
This whole article is based on a completely wrong premise: That the computer of a self driving car would be programmed or able to predict that people would be injured or killed. It doesn't work that way.
What the computer cares about is hitting things. It avoids hitting things. It will likely be programmed to give preference to avoiding soft things, a bit more preference to avoiding human-looking things, and maybe a bit more preference to avoiding small human-looking things (but that depends on whether small human-looking things suffer more damage from a collision than bigger ones, which I don't know). Very high priority would be given to avoiding hard and heavy objects moving at speed in the opposite direction (oncoming traffic, especially trucks or buses), and a lot less to objects moving into the same direction (like a bus driving in the same direction).
A future self-driving car will be better at protecting you, because if there is an accident, it can predict it ("I will crash into this wall in exactly 0.105 seconds"). Steering wheels and pedals can be moved out of the way, seat belts tightened up, airbags exploded at exactly the right time - not waiting until the crash happens.
Now the scenario of ten pedestrians in the street... That would happen in an area with a speed limit. Whatever happens, the car will be braking from 30mph to 20mph or lower, which with preparation is quite safe. Even if there was a situation where the car drove into oncoming traffic, with lots of cars computer controlled a message would be sent out so that everyone brakes hard and gets out of the way. One human-driven car driving into oncoming traffic is not the same as one computer-driven car giving its mates a warning long ahead, everyone does an emergency brake immediately, and then you go into incoming traffic.
Proving that Peter Dizikes is an amoral, selfish bastard, it seems to me.
I would have said "the answer depends on whether the crowd is composed of Westboro Baptist protestors and child-molesting Catholic clergy, or a bunch of pretty girls holding kittens" - thus proving what a moralizing, vindictive bastard I am.
If you only analyze and react to situations based on your personal gain, you can (and probably should be) replaced by an Ayn Rand ® brand robot. Humans are more complex than that, and will cheerfully give up their lives for ideals or even whims.
I have just two words for you: "Facial Recognition". And a quick DNA analysis of course, before the crash.
"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."
I don't. Not at all costs. I don't drive that way and I wouldn't want others to, either, whether in a self-driven car or a self-driving car.
Take a hypothetical example. I'm driving down the street, observing all relevant traffic rules when an oncoming vehicle, for no apparent reason, swerves into my lane and is headed for a head-on collision with me. For the sake of this hypothetical, I have only two options, as in the above example-- to either allow the collision or swerve myself. The only available location to avoid a collision is occupied by multiple pedestrians who are also obeying the traffic rules.
Plenty of people would swerve reflexively and hit the pedestrians. If law enforcement and the legal system work properly, I would most likely be exonerated in this accident, and blame for any damage or injury suffered both by myself and the pedestrians would be borne by the driver of the other vehicle.
Which does nobody a damn bit of good if the pedestrians are dead, and possibly no one left to blame (except an insurance company) if the driver of the other vehicle also died in the collision. (Let's presume that if I avoid him, he hits the car behind me, killing that driver and himself. Just for the sake of argument.)
A human driver can defend themselves by saying they didn't mean to hit the pedestrians, perhaps that they didn't see the pedestrians prior to the accident; if they are not in a crosswalk and not competing with the car for right-of-way then there's no particular reason to have observed them.
The autonomous car has no such excuse. It might fail to avoid the oncoming car, it might fail to avoid hitting the pedestrians, but it will not be because it had no time to consider the possibilities or because it was frozen by indecision. I really don't think there's any way to approach this other than a utilitarian one. The car can't be expected to accurately compare the relative risks of damage, injury or death to the various parties. It can probably, however, be supplied with information about the number of occupants in the car and the number of pedestrians on the street.
I would hope I would not swerve into a dozen schoolchildren to save myself from an oncoming vehicle that is violating the regulations (for whatever reason) just to save myself and transfer the damage for which that vehicle's driver is responsible onto somebody else. If I can safely avoid him without significant involvement to bystanders, I should. If I can't, I should not. I should let them hit me, restricting the involvement in the accident, if possible, to the perpetrator and the one vehicle our example assumes will contain victims: mine.
Now, should my car hit other cars to save me? Presumably unoccupied, parked cars, for instance? Sure! Property alongside the road? Probably! Utility poles? Well, now we're getting into an area where the car won't and can't know what would happen, so probably the car should prioritize avoiding pedestrians and then structures like utility poles, but pay less regard to other cars, either stationary or moving.
I did read the comment and the comment before it that talks that the situation is on a highway
Slowing down just because you are going under is something no driver would do when you are in a highway\high speed road, if the speed is 100 km/h why would you slow down to a reasonable speed, like 40 km/h?
You're assuming we all agree Nazi is bad.
The car logic would be able to predict such a condition and slow down way before this scenario would happen in the first place rendering this whole line of thinking moot.
In the event that it happens anyway, go ahead and run 'em over. If they were stupid enough to get into the street in the first place, they deserve it. Darwin in action.
Hopefully, the driving ability of AIs will improve to the point where cars will no longer need to be equipped with expensive airbags. (Yes, you would have to get almost all of the human-driven cars off the road first, to get rid of the threat that they pose.)
That that is is that that that that is not is not.
A choice between saving 10 lives and saving one life? Why does anyone consider that to be a dilemma? It's a no-brainer.
Kill the one to save the 10, for a net savings of 9 lives. If any machine is programmed to do the opposite (kill 10 to save one, for a net loss of 9 lives), we should seriously think about bringing criminal charges against the programmer.
That that is is that that that that is not is not.
If they ever get to the point where the auto-drive cars can even recognise the dilemma, then it will be valid to argue about it.
But when humans can't figure it out in their easy chairs, much less in their cars, don't expect the computers to do it.
Computers are not superman and they are not gods. They will end up even slower than people to respond in such situations.
Just look at what the software load did to PCs, over the years. PCs are orders of magnitude faster, but the software has so much added to it that they end up slower than before. It will happen to the cars, plus the government beauracratic requirements! 8-P
Therer are always more than two choices. Just because humans have only two hands, doesn't mean the world is that way.
Think with your feet! Find additional choices! ;-)