Slashdot Mirror
Who Is Liable When a Self-Driving Car Crashes?
innocent_white_lamb writes "Current laws make the driver of a car responsible for any mayhem caused by that vehicle. But what happens when there is no driver? This article argues that the dream of a self-driving car is futile since the law requires that the driver is responsible for the operation of the vehicle. Therefore, even if a car is self-driving, you as the driver must stay alert and pay attention. No texting, no reading, no snoozing. So what's the point of a self-driving car if you can't relax or do something else while 'driving?'"
A self driving car would be less likely to rubberneck, or cause other issues relating to a human driver. Cars could in theory go faster. etc.
I would think the point would be that machines, once properly programmed, can be the worlds safest drivers...statistically. You, as a human, will still be responsible for taking over when the machine doesn't know what to do. But, for the other 99.5% of the time, the self-driving car will make the best decisions and always be completely alert.
Self-Driving cars, I believe, have the ability to drastically reduce deaths caused by motor vehicle accidents...one of the highest causes of death in the USA.
There's an industry that manages risk.
Regulation (e.g., insurance) always develops spontaneously, because there is a market for reducing chaos.
If neither party is engaged in any such activity, then an age based system will be utilized.
Current law not appropriate for future technology! News at 11!
Talk about a crazy-assed prognostication! This is a ridiculously stupid question (cue the "even by slashdicetimmy standards" responses).
you might as well ask what would happen if it turned out that the number of angels that can dance on a pin turned out to be finite.
https://app.box.com/WitthoftResume Code: https://github.com/cellocgw
If the car has a software issue and crashes then the software developer is at fault. If the car has a hardware problem then the hardware developer is at fault. If the car has a mechanical failure then then mechanical engineer is at fault and so on. Either developer the components / modules correctly in the first place or not at all. If modules / components have lifespans then just lock the car from starting once those lifespans have been reached and if you don't want to be held holding the torch when shit hits the fan then don't get involved from the get go. To spite this modern system of pass the buck and never accept ownership of the problem, someone caused the issue by not doing there job right to begin with and they should have to rectify it.
A car that drives itself is responsible for itself.
Who pays in the event of an accident is the driver. In this case, the car. Probably the manufacturer would be liable.
Manufacturers will probably get insurance for the car when driven autonomously. If self-driving cars are safer, this should be a lower insurance rate than you pay now. Additionally, self-driving cars will probably have sensor input that will prove/disprove fault.
-- Erich
Slashdot reader since 1997
Maybe it should be like govt caused problems, where the taxpayers pay for all problems.
But with nothing to do behind the wheel 99% of the time your not going to be alert. Your going to be super bored. So when your supposed to take over you won't be prepared to do so.
The manufacturer will have an EULA which absolves them from guilt.
It won't be the people who sold it, because they'll also have a contract term which says they are absolved from guilt.
So, it will come down to the owner, who will be entirely dependent on the quality of the product, as delivered by two entities who have already said "not us".
So, if you privately buy an autonomous car, and it crashes, you will likely be on the hook for it. If you merely hire them (as in a Taxi), then I'm sure the people who rent them will also absolve themselves from guilt in some strange way -- likely through arms length 3rd parties who do the actual operation.
This won't be so much "buyer beware" as "everyone else on the roadway beware", because you'll have a vehicle driving around that if it crashes, there's a long line of people who have already made sure their asses are covered.
The lawyers for the companies making and selling these will have covered their asses before it ends up in the hands of anybody else.
Lost at C:>. Found at C.
A vehicle malfunction that causes an auto accident won't be attributed to the driver. When Toyota's gas pedals were getting stuck and causing deaths, the lawyers were going after drivers. There's no difference with autonomous vehicles. If the technology is found to be at fault, it will be the part manufacturer and the auto maker who will be dragged into civil court.
Just from memory:
Montreal Metro is driven by autopiloting with someone in the cab for door management.
Vancouver Skyline doesn't even have a driver anywhere, it's all automated.
Several airports (Orlando was the last one I went to), have automated trains/monorails to shuffle people between terminals.
Most flights you take are done almost entirely on autopilot.
So far, it seems that mass transit is increasingly automated. So why is non-mass transit any different?
I'm god, but it's a bit of a drag really...
What? Simply wrong. You cause an accident, you pay. If your brakes fail, after you pay, you might have a civil case against your mechanic or car manufacturer.
John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
Open your eyes to the more distant future where all new cars are self-driving and only antiques require drivers. Then we can lay blame of self-driving-car-on-self-driving-car accidents with the manufacturers.
It would be the exact same liability that it is now: the manufacturer is at fault for manufacturing defaults, software or otherwise.
This isn't an issue.
The Kruger Dunning explains most post on
And the liability will shift to the manufacturer of the autonomous vehicle more so than the person riding in it and owning it.
The change will happen slowly, organically, over time. A self driving car will behave statistically as a very safe driver. Ownership of a self driving car should bestow upon you lower insurance rates. If the current insurance companies balk at the idea, the private market will take over and "self driving only" insurance companies will gladly take their place. Eventually, as more and more share of vehicles are self driving the size of the insurance industry will shrink significantly.
I see no reason to change the liability burden away from the "Driver". It may seem counter intuitive, but you are gaining economic advantage by using your self driving car. For that advantage, you accept the risks, and insure yourself against them. That said, operating a self driving car will/should carry significantly less risk and liability then driving yourself around does now.
That does not mean that the car makers are off the hook. Just like today, if a vehicle mechanically malfunctions in a way that the car maker is found responsible, the insurance company may attempt to subrogate the claim to them.
Remember the toyota software bug? Toyota cars had a software bug that caused older drivers to accidentally hit the gas when they wanted to hit the brake! But it only affected older drivers, and driver height was also a factor. Anyhow, an expert witness reviewed the source code at testified at a civil trial that he couldn't find any bugs but he couldn't rule out that bugs existed. The jury found toyota liable. Cha-ching!
So even if you use something like haskall to prove your code correct or node to prevent blocking, the car manufacturer may still be responsible for drivers pressing the wrong pedal. It's software, there might be bugs, pay up.
Do you even lift?
These aren't the 'roids you're looking for.
Who is liable if you have a crash in a taxi cab or a state-owned vehicle? The thing this article overlooked is that there is more than one business model for selling cars. Self-driving cars might flourish by allowing companies to provide a lower cost car service for those who either cannot or do not wish to drive themselves. Apps like Sidecar (http://www.side.cr/safety) and Lyft (http://www.lyft.me/safety) are already pointing in this direction and centrally controlled driverless car services could be a logical next step, especially if companies take on the liability for what happens during a ride -- just as they would in an airline, rideshare or taxi service.
Moreover, even if driverless cars don't become the norm, driver-assist cars may do so and could dramatically reduce accident rates. As a car and driving enthusiast, I am selfishly averse to all these changes, but the safety benefits are hard to argue against.
The first principle is that you must not fool yourself - and you are the easiest person to fool. -Richard Feynman
Just make the car white... and put a fruit symbol on it. Millions of people will buy it despite the fact it has no practical application.
In many areas, this is not regulated by law but by legal precedent. Besides, laws can be changed and precedents evolve.
Besides, depending on the cause of the accident, this could easily fall under existing product liability laws, regulations, and precedents.
...there has to be *somebody* who can be sued. It's the American Way.
Some sort of no-fault insurance that all driverless car owners would pay into that accepts responsibility for and pays out damages on accidents seems like the obvious solution here.
If the cars are genuinely significantly safer than it would be cheaper than current insurance. And if there is an accident, the damages are covered, and there's no penalty to the owner.
This doesn't seem like an intractable problem at all.
The driver is not the [only] one liable.
First, any person, even a pedestrian or passenger who causes an accident through their own acts can be held liable. Even if you swerve to avoid a dog or child who runs into the road and sideswipe the car next to you, the parent of the child or owner of the dog will be liable for the damages (unless, of course, you failed your duty to keep a proper lookout or otherwise acted negligently).
Second, the owner of the vehicle is also liable, for act of someone he authorizes to use the vehicle.
Third, if the car was in use on behalf of an employee/servant the employer/baster is responsible.
Fourth, the manufacturer is liable if a product defect is responsible.
Finally, none are liable if they (including their property (i.e. car) and their agents/employees/servants) were not the proximate cause of the accident.
I was researching the appropriate statues in the Combined Annotated Statues of the Law of the State of (wherever) at the time the vehicle ran down six members of the State Supreme Court. I refer you to Evidence Photo #17, in which the rest of the car was full of lawbooks. your honor, this case should be considered pre-appealed, as it has already been presented to the Supreme Court, and I should be released on personal recognizance... .
if this is supposed to be a new economy, how come they still want my old fashioned money?
Make it so the company who makes the car is "the driver" legally.
In other words, don't sell us your shitty cars if they crash.
Generally it is the rider. I have seen drunk riders pulled over in New Mexico, where horses are used when a driver has their license taken away for DUI. Seriously, I have seen this twice in Taos.
Then I guess they get charged with RUI.
If your breaks break, does that mean they stay in one piece?
It's because of this conundrum that autonomous vehicles will only be novelty features on standard automobiles. It will be an auto-pilot or cruise control wherein the driver is still expected to take control in the case of an emergency that could not be measured by the car's sensors or accounted for by the car's algorithms.
And that's not bad! It's just not as idyllic as some would prefer.
Drivers are going to be alert for about two weeks and then the novelty and thrill will have worn off, and he'll be like the guy who works at the amusement part on the roller-coaster. Yawn...
If you want to read, or nap or do anything other than pay attention to driving just use public transit. It's not always an option, but if you really just don't want to worry about driving it's the best choice. And it adds efficiency that even a self-driving car can't bring to commuting.
Even if I knew that tomorrow the world would go to pieces, I would still plant my apple tree. -Martin Luther
All one has to do is compare it to a situation involving another driverless vehicle: a car that rolls down a hill. If one owns a car parked on a hill and, for whatever reason, the car rolls down the hill, one is liable for any damages that results as well as any fines or penalties.
If one believes the car rolled down the hill because of a defect in the car, then one can attempt to hold the manufacturer of the car liable for the damages, penalties, fines, etc..
If one can show that a third party did something to the car to cause it to roll down the hill, then one can attempt to hold the third party liable.
In any of the cases, the owner of the vehicle is primarily liable. In the second and third cases, the owner can attempt to recoup losses by holding a third party liable, but will still be liable for any damages caused regardless of resolution with the third party.
There is no "-1 offended" or "-1 you don't agree with me" mod options for a reason.
Whose insurance company, the car's owner, the car's driver/occupant (the person who put the car in self driving mode) or Google's (the self driving software developer)?
Insurance companies are not liable, they just pay the bill for the person or company who is liable.
so you sit in jail / prsion after a software fault ends up running a kid over?
I see two cases:
1. The car is designed to be semi-autonamois but monitored by a driver. In this case the driver is responsible for it juts like they would be if they failed to stop in time wile using cruise control.
2. The car is designed to be fully autonomous and run without a driver. In this case the car's owner is responsible for its actions, much like how an owner is responsible for the actions taken by their pets when unattended.
Let's say the software is very buggy and they hide that, withheld updates, and so on.
Most financial responsibility laws specify a very low (say $50,000) liability coverage requirement. That is about 100 times less than what you can get if you are killed in an airline crash.
If a self-driving car kills you and you can sue Google (or whoever), your heirs will get several million dollar dollars, instead of $50,000. In other words, until a self driving car has an error-rate 100 times lower than humans, they won't be made.
"When my accounting software gets it wrong, they eat the bill."
You may want to read the find print, depending on the software you use. Most of them say something along the lines of giving your free audit support but you're still liable for any fees/penalties owed. It's also not a liability thing that they offer, but an added feature to get you to buy their software.
The insurance company is not liable, they merely pay the bill for the person or entity that is liable. Google, or whoever develops the software, will need insurance just as much as the driver since they may be found liable if there are software defects.
People already accept full liability for their car, despite texting, reading, snoozing, having sex, playing games or any of the other horrific distractions that far too many people engage in. So the easy answer is that people will continue to accept full liability for their vehicle, and may optionally take out additional insurance for accidents that occur while their car is in auto-drive. The incidence of accidents for auto-driven cars should be low enough that their normal insurance premium is much lower, thus additional insurance specifically for accidents that occur through automation failure plus human inattention should also be affordable.
So:
* Montreal Metro -- autopilot with a trained person there to monitor, just as noted here. Also, on separate tracks with no route diversions to worry about, and no other traffic to contend with.
* Vancouver SkyTrain -- autopilot on separate tracks with no route diversions to worry about, and no other traffic to contend with
* Airport trains / monorails -- autopilot on separate tracks with no route diversions to worry about, and no other traffic to contend with
* Airline flights -- autopilot and sometimes autolanding, monitored by at least two highly-trained individuals. Route diversions and other traffic do exist, and modern aircraft will take diversions automatically as a last resort, but these systems are still manually configured by the pilots in different phases of flight, and usually the pilots take action long before they come into play.
So really, none of these examples are in the least bit relevant.
There's a straightforward solution. Lease autonomous vehicles on an operating lease, with insurance and maintenance included. I used to get a deal like that when I was a Ford employee, years ago. The manufacturer is partly in the insurance business; they cover little claims directly, and reinsure against big ones. Since they're reinsuring many cars, they get wholesale rates.
Now a single party is responsible regardless of whether the driver or the hardware is at fault. They can sort it out internally. Also, if maintenance is included, the manufacturer has control over maintenance quality, so they get to recheck the autonomous driving components on each scheduled service.
That's enough to handle the first stage of deployment. Once insurance companies see the automakers moving into their business area, they'll offer competitive rates.
I have written about this several times now. This is what is going to happen in the U.S. It is as predictable as sunrise:
A "self-driven" vehicle will be in an accident. Maybe it kills a passenger, or somebody else. This might even happen several times. The hardware/software will be blamed. There will be big lawsuits.
The "self-driven vehicle" craze will die down for a while. People will say "I don't want to go near one of those things."
Eventually -- and it may take quite a while -- the technology will improve beyond question and they will be adopted again.
It's happened with just about every major advancement in automation to date. Why should cars be any different?
IAAL.
Under current law in most states, the owner has financial responsibility if there is a crash. In fact, most car owners have no assets to collect on, so it is the owner's insurer that usually picks up the tab.
Regardless of manufacture attempts to disclaim liability, third parties (drivers and passengers of other vehicles, pedestrians) can sue the manufacturer if the crash was caused by an unreasonably defective vehicle. It is relatively rare because most cars are insured and the owners insurance covers the injury, while claims against the manufacturer are expensive and difficult to prove.
Further, anyone who thinks that there are not driverless cars on the road today is kidding themselves. I see many driverless cars on the road. We will be far better off when there is computer operating the car instead of an idiot texting behind the wheel.
In the land of the blind, the one-eyed man is king.
"Therefore, even if a car is self-driving, you as the driver must stay alert and pay attention. No texting, no reading, no snoozing. So what's the point of a self-driving car if you can't relax or do something else while 'driving?'"
The same rules apply today and the roads are still full of drivers doing anything but paying attention to the task at hand. Those same people will happily let a computer take on the last vestiges of their responsibility and zone out completely, then express shock and ignorance when something goes wrong.
Figuring out liability for human drivers is insanely complicated. You just don't notice it because drivers are generally removed from the issue; there is a whole industry (car insurance) whose existence depends on profiting from driver liability and thus deals with all of the complications for you. They even figure out things like risks of being hit by an uninsured driver and factor that into the cost. The only cases they don't handle are when 2 uninsured motorists get into an accident, and then the courts can get involved.
Insurance companies will figure out the risks of various types of autonomous car failures and to what extent their liability costs can be recouped from the manufacturers (due to negligence), from the passengers, from the other parties involved, etc. Then they will set their insurance rates for autonomous cars so that they can cover liabilities and still make a profit. If they underestimate their liabilities then they raise their rates or go out of business. The autonomous cars could even require proof of insurance to be installed or downloaded in order to operate, making them very difficult or impossible to operate without proof of coverage. That will pretty much eliminate the problem of uninsured drivers.
The only thing that would hold back autonomous cars is if the risks are estimated to be too high, making the insurance rates so expensive that it outweighs the convenience. Given how unsafe most human drivers are, I think the autonomous car manufacturers would have to do a really terrible job for that to become an issue.
The owner/operator of the car. That's who has to settle first. If it's defective hardware then the owner operator needs to file a claim against the manufacturer. The same goes for the software.
I'll buy a self-driving car when it comes with a full bar, and an in-dash ice machine.
Works very well in some areas. Basically nullifies this article.
Who is responsible?
Driver gets in the back seat with his honey (or kids) and the other kid gets bored & gets behind the wheel and "takes over."
Hey we have a lot of lawyers out of work.
Have you ever looked at the fine print of the licensing for Microsoft Windows?
It clearly spells out the software isn't suitable for health/medical devices, aircraft maintenance, and a whole raft of things.
But, there's tons of stuff which do those things using Windows.
But since you've agreed to the license terms (you're using it after all), the license terms say "we bear no responsibility". And since license terms have been upheld, you can safely bet than anybody making autonomous cars will have also covered their asses.
And, then of course, there's the fact that they bribe politicians into passing laws in their favor by using lobbyists -- whose job it is to ensure they carry little of the risk.
You'd likely have to demonstrate some pretty widespread malfeasance to actually hold them responsible for anything. But, since Obama was happy to give Monsanto retroactive immunity, I'm sure some other politician will also do this.
In other words, assume they're buggy as hell, dangerous, and simply don't buy one. Because that's what I'm gonna do.
Lost at C:>. Found at C.
Florida (and maybe some other states) have no-fault car insurance. They don't try to figure out who is at fault in an accident, everybody's insurance policy covers damage their own vehicles, regardless of who caused the accident.
This principle could be applied to robot cars as well.
I don't see why the model should change. The driver of the car would be liable for any accidents caused by their car. This will probably be covered by the driver's insurance. The insurance companies will only insure self driving cars that they consider safe. Insurance companies my even offer cheaper rates to cars with a better track record of self driving.
Maybe the cars could have AI personalities, like cautious, aggressive, or FastNFurious. You can set the AI based on whether or not you're in a hurry or how you feel that day.
It'll be like aviation - the makers of the craft in question will pay lots of mney to lawyers to put the crash down to "pilot error".
Aside from which, let's see, law enforcement will want a 'kill switch' and every politician will want a 'Zil lane' button.
All your ghosts are just false positives.
The parent would be responsible for the actions of their minor children under their supervision. If the parent is not the owner of the insurance policy of the vehicle, the policy owner, or the insurance company itself, may attempt to recover damages from the parent.
Obviously each case is different, and lawyers will have some work. Your hypothetical situation is similar to the "brake shift interlock" issue that went through the courts a few years ago. Parents argued that it was unsafe to be able to shift a vehicle out of park without depressing the brake after several children were injured and caused damage by doing so. I believe the interlock is now a mandated safety device.
I think people are over complicating this to death. It doesn't matter how magic the technology is. Insurance is black, cold, flat risk assessment. Nothing more. Everything else is details, and the lawyers will _continue_ to make plenty of money on that whether we ride horses, or have automated flying cars.
No, just as one doesn't sit in jail if one runs over a kid because one's brakes fail for a reason that one is not responsible for, such as damage to a brake line caused by road debris. But, if one hits another car, one is still financially liable for the damage to the other car and can possibly attempt to recoup the money by civil suit if a responsible party can be determined.
There is no "-1 offended" or "-1 you don't agree with me" mod options for a reason.
There are two distinct things: One is that you are officially the driver even if the car drives itself, and you are responsible. But the whole point of a self driving car is that it is safer driving in a self-driving car with your eyes closed than in a non-self driving car with open eyes. You are responsible, but nobody is going to say "you are responsible because you used a self driving car without watching". They will say "you are responsible because your self-driving car caused the crash". Which will happen less often than if you drove yourself.
Right now you have to (a) watch out what you are doing and (b) pray that you don't have an accident. With a self driving car you don't need to watch out what you or the car are doing; you still have to pray that you don't have an accident.
And the whole idea of taking control in unexpected situations is nonsense. In the very best case, you would have to (1) do something to take control away from the computer and (2) react to the problem. In situations where there is enough time for that, the computer can handle things just fine. And people may think they are good in unexpected situations, but they are not.
Again, I ask, Why should the owner of a car pay insurance, when they are not in control of the vehicle?
Sig Battery depleted. Reverting to safe mode.
Its a good thing you are posting as AC.
Sig Battery depleted. Reverting to safe mode.
That describes criminal negligence. The owner of the vehicle is still liable for the financial costs of the wreck, but can file suit against the manufacturer to recoup losses.
There is no "-1 offended" or "-1 you don't agree with me" mod options for a reason.
And, Microsoft would be in clear but the company that build the software would be liable for, possibly criminal, negligence for using an inappropriate operating system, regardless of the EULA of the car/software.
There is no "-1 offended" or "-1 you don't agree with me" mod options for a reason.
Not that I'd like it, but I'm guessing insurance companies will switch to selling 'normal' insurance that is only valid if/when you have your car in self-driving mode.
If you want to be covered all the time it will cost you a lot more.
Conditional insurance is somewhat similar to what they already do in Germany. Even if it is legal to go > 90mph on the autobahn, in most cases your insurance becomes invalid > 90Mph.
Generally it is the rider. I have seen drunk riders pulled over in New Mexico, where horses are used when a driver has their license taken away for DUI. Seriously, I have seen this twice in Taos.
Then I guess they get charged with RUI.
There is no such provision in New Mexico law.
http://law.justia.com/codes/new-mexico/2011/chapter66/article8/section66-8-102
Unless the horse was pulling a wagon, it is not a vehicle.
You might be charged with drunk and disorderly citations, but you will not be charged with RUI.
Its equally possible you just made this up.
Sig Battery depleted. Reverting to safe mode.
When he's ready, he won't have to.
AJ Henderson
While the actual numbers vary by car and conditions... The example in the chart shows roughly 33mpg at 50mph, and 20mpg at 80mph, but doesn't consider travel time.
You can decide if you value your time at more or less than 2.6 gallons per hour.
This issue is a bit more complicated than you think.
Again, I ask, Why should the owner of a car pay insurance, when they are not in control of the vehicle?
You may own a self-not driving vehicle aka a house. You don't control where it goes; it's not supposed to go anywhere. If a roof tile drops off your roof and injures me, you are liable.
The owner of the car would be liable for the cost of an accident caused by the car because they are the owner of the car, and they are clearly in control of whether the vehicle drives or not. The owner of the car pays insurance because society demands that car owners pay insurance so victims won't get stuck with the cost.
Otherwise, what's the point in buying a self-driving car? If I have to monitor it every second to make sure it's not crashing, why even bother? People bring up aviation, but it's not the same, the pilot is always involved to some extent, better trained than most drivers, and if there's a problem with the auto pilot there's usually plenty of time for the pilot to take over and make corrections. Not so on a crowded roadway.
Obviously they can't build the entire cost into the purchase price, so I'd think they'd give one year free coverage and then charge the customer beyond that. Of course, that would depend on all systems being functional, so maintenance becomes an issue. I expect at first these cars will only be leased so manufacturers can keep tight control of them.
Never let a lack of data get in the way of a good rant.
Because as an American living in a multi-cultural society I do not want to share my resources with anyone who doesn't look like me, think like me, vote like me, or has a dissimilar amount of purchasing power. I prefer to associate only with those of my own socio-economic-ethnic background, because I'm sick and tired of having to pretend to get along with everyone else. I only do that for my own family, mind you. Everyone else is weird, stupid, and they all smell funny. And then you're asking me to subsidize the poor driving habits and behaviors of those who are inferior to me (or whom I believe are inferior, because obviously they -are- inferior to me)? And even worse, you're asking me to subsidize those who are superior to me (those damn bastards, they already have the world) and their fast hot sexy cars? Fuck them, and fuck diversity--I want my own damn monoculture because I'm tired of being uncomfortable around those who are different from me. So I want insurance and health care tailored to my particular socio-economic-ethnic background and damn all the rest to hell.
Here's to hot beer, cold women, and Glaswegian kisses for all.
Because the software and hardware (and maintenance costs) for a self-driving car is far (FAR) beyond the reach of even above-average wage earners, it will likely be rammed down our throats by insurance companies, and then funded by privacy-invading data collection, and . . . wait, you're not busy paying attention to the ROAD while you're driving? We'll show you ads. And we'll select a route that drives you past billboards, and shopping malls. Oh, you can opt-out of that if you want. For a "small" fee.
Expect lots of lobbyist-driven legislation to support this model.
These are my friends, See how they glisten. See this one shine, how he smiles in the light.
When a surgeon is first out of medical school, they will do dozens if not hundreds of surgeries under the direct supervision of an attending surgeon. If that med student starts to do something dangerous, the attending (who was there the whole time watching) will step in and help them out. Any time the master steps in the student learns something new, effectively making it better. This doesn't mean the attending will lose focus after a few minutes and start reading a paper or playing on his/her phone, it just means they aren't the one actively performing the surgery. If the student destroys someone's knee or kidney or bowel the attending is held responsible, but the student gets bad marks for causing the situation in the first place. It is an effective feedback system.
Brad
If a roof tile drops on you, its only because I didn't maintain the house. I have control of my house.
If a tornado blows it off, its not my fault, (act of god) and your own insurance handles it.
Sig Battery depleted. Reverting to safe mode.
So does this mean the horse can be drunk?
This meme of "self-driving cars will never work, because who gets sued?" keeps popping up, yet the idea of having liability insurance for personal possessions not under your direct control has been around for a long, long time. If someone visits your home and hurts himself while on your property, your homeowner's liability insurance covers you, even if you are not physically present. The insurance companies will learn to deal with self-driving vehicles, because there will be money to be made, and they will figure out a way to get into that market.
In any case, self-driving cars are absolutely inevitable for one major reason: our aging population. Senior citizens are going to demand the freedom of personal transportation, and anyone in the U.S. who tries to tell them "no" is going to be fighting the AARP, which has some of the most powerful lobbyists in Washington. Furthermore, consider citizens who are blind, or deaf, or epileptic. Why shouldn't they have the right to personal transportation? This will become a mandate for individual rights enforced by the federal government.
In any case, people who claim self-driving cars will never work keep ignoring the elephant in the room: 35,000 fatalities and 2.2 million injuries a year, and a cost of $250B due to car crashes - and that is just in the U.S. alone. We slaughter each other right and left, and just shrug our shoulders. I'd much rather trust a computerized driving system, even if it has rare failures, because statistically I'll still be much, much safer on the road.
Ultimately, this argument will all be moot. It reminds me very much about how some people railed against personal cell phones when they first began to appear. How did that work out? In thirty years, you'll have a whole generation of adults who have grown up without having spent 5 minutes of their lives behind the wheel. At that point self-driving cars win by default, because most people won't even know how to drive anymore. To them, knowing how to drive a car will be about as relevant as knowing how to saddle and ride a horse.
You're liable for things that happen in your home, but you don't have to stay home all day and be vigilant about every little thing that happens. That's what insurance is for.
Higher Logics: where programming meets science.
driverless cars.
Insurance companies will only insure driverless cars if the accident projections prove true. But wait, where's the data to start up? They will estimate using various statistical techniques in which their actuaries are well-versed to determine if it is possible to turn a profit based on the limited test data available. If they decide it is possible, they will then invest in their political lobbies to get some laws written that will provide large government subsidies for them to take on this extra "risk"- i.e. they will transfer the risk to tax-payers. As the subsidies dry up, they will accumulate data and adjust their models as more and more driverless cars hit the streets. Then they will figure out how to transfer financial responsibility back to the "driver" or manufacturer and, using the healthcare insurance model, will charge ever increasing premiums while delivering ever decreasing coverage.
It's the American way! It's what American "exceptionalism" is all about!
That's a distinction without a difference.
No. Insurance policies often have limits. If you have a $1M policy and you are found liable for a $5M judgement then your insurance pays $1M and you are personally responsible for $4M.
Most people already own and 'control' some automated systems, and we have a functioning legal framework for establshing liability. Some examples:
1. An automated sprinkler system. If your sprinkler system sprays outside of your property onto some passerby's extremely expensive suit, you as the owner of the system are liable. You are in control of the system, have set it up to spray beyond the bounds of your property, and have given the system instructions on when to start and stop spraying, usualy via a timer. You could argue that the system is deficient, or there is a manufacturing fault which prevents the system from operating as intended, i.e. spraying where you tell it to within commmonly acceptable limits. If the nozzle comes broken and it sprays 360 degress instead of 60 degrees, you can attempt to pass the liability for the incident on to the manufacturer, assuming you can convince people that you didn't know about the defect in advance, in which case you should have acted to mitigate it.
2. An automatic robot vacuum cleaner. I have one, and while it's not a common thing in households yet, the liability surrounding it is pretty clear. You can block of certain areas with a magnetic strip that it comes with (it's a neato). In summer it's really hot, and I like to leave my front door open to get a breeze going through the house, but the neato can just fit under the security gate. So I put the magnetic strip across the front door. If I forget to put it there, and one day it goes off to clean in the street outside and it cause an accident, I'm liable, not Neato (the company). It's functioning as intended. I accept this risk as part of owning the robot. If it decides to go AWOL and *cross* the magnetic strip, not performing as intended, and cause an accident in the street, then Neato (the company) is responsible, or at least I have a good case to make them responsible.
Now why wouldn't an autonomous vehicle fit into the same framework. If an autonomous vehicle causes an accident while not under autonomous control, it isi definitely the drivers responsoibility. If an accident happens while the vehicle is under autonomous control and functioning as intended, it's the passenger's responsibility, because they presumably gave the instructions to the car, however since the instructions to the vehicle would be very high level, as in 'take me to X destination' and the intended function of the vehicle is to get you to that destination safely, you would have to have issude an order like "take me to X latitude, Y longitude" which happens to be just the wrong side of a cliff. Even then, the vehicle *should* avoid situations like these. If the vehicle does not perfrom as intended, i.e. it collides with something, which it is not meant to do, then it is the manufacturer's fault, or rather, again you have a good chance of making it their responsibility assuming you can prove the car would act consistenly incorrectly in the same situation. As far as disclaiming liability is concerned, make the autonomous systems get the euivalent driver's licenses, i.e. to sell a driverless car, it must be certified as capable of driving autonomously, which could be made to mean: will always act to avoid a collision in all directions, will drive on a particular side of the road, will negotiate with nearby autonomous vehicles using a standard protocol, will obey traffic signals, etc...
The legalities are not that hard!
no-fault insurance.
Shoes for Industry. Shoes for the Dead.
If we don't want an even worse repeat of the 80s, when Germany and Japan were taking over the auto industry, our laws will have to be made friendly to autonomous cars. The countries with more inviting laws will be the ones who perform the most development in the early years of this technology. Then once the technology is "solved" and the entire world is open to autonomous cars, those countries who were early adopters will be the ones running the industry. They will already have the expertise that other countries don't have because there wasn't a local market for the cars yet.
-- All that is necessary for the triumph of evil is that good men do nothing. -- Edmund Burke
You will turn you car over to the "hive mind" to get lower tolls on the expressways. In addition you are likely to get higher speeds, lower gas cost and greater safety. The cowboys will still have non-automated lanes to try to drive faster, but that may be unlikely.
A self driving car is that much less likely to be stupid. It's not going to jump lanes in traffic because it's in a hurry, or because it has a self-delusion of being the next Mario Andretti. What it will do is adapt to changing driving conditions, such as rain, ice, etc. It's not going to succumb to highway hypnosis, even if the human "driver" falls asleep at the wheel. The decisions it makes will be all derived from logical algorithms. The challenge is to make those algorithms smart and adaptable enough. Sure there will be accidents caused by those algorithims, but most likely far far less than those caused by human error.
And if the car caused the accident because of a manufacturer defect (including software) the insurance company has enough heft to go after the manufacturer (which the driver probably doesn't).
As you automate for decisions for one reason or another, of machine that coukld harm humans, then the questionof liabilility becomes an issue. Medical assistant machines, like cancer radiation machines, have been in the news. Sometimes their programming has been in error despite a lot of preventive development. You cant just write it all off with a disclaimer.
"Don't say 'murder,' don't say 'kill' It was destiny, it was God's will"
See more here
while [ 1 ]; do echo -n -e "\xe2\x95\xb$((($RANDOM&1)+1))"; done
No, it's not solved. Insurance company claim payout comes from the (collective) car owner's pocket. Why should the car owner be liable for accidents he did not cause?
Google's team announced that they had passed the 300,000 autonomous mile mark on public roads. Accident-free.
While super impressive that may not be enough miles to be statistically significant for comparison's sake. I can't find all accidents but for fatal accidents the accident rate in the US is around 1.5 fatal accidents per 100 million miles driven. So far they are at 0.3% of 100 million miles so they really need to drive a lot more before we jump to serious conclusions about comparing safety rates.
I'm curious what the total accident rate per mile driven in the US is. Would be interesting to compare as more autonomous miles are driven.
I'm not buying a self-driving car until I can sit in the back seat and drink a beer.
Kid-proof tablet..
If the promise of self-driving cars includes the idea that the self-driving car will rarely ever be in error and will at least be far less likely to be in error than a human driver, then it seems probable to me that a human being being alert enough and able to correct the self-driving car may make things worse by trying to intercede both legally and in terms of actual outcome.
Also, market forces might cause insurance companies to offer lower rates for cars that are self-driving, and eventually much lower rates, because they know they will almost never have to fork over the money and they want that market. At some point, courts and the general public will figure out that it was almost certainly the fault of a human-driven car and as such, liability may end up being nearly always on the human driven car, driving up insurance for cars built for or intended to be used by humans drivers.
Human drivers might end up being priced out by the rising costs of gas (self-driving cars are probably going to be more economical), liability, and so forth. Once a car can really be self-driving, we can have probably pretty damn cheap self-driving taxis and "minivans" which use algorithms to pickup multiple people in a small area who want to go to a similar place, further driving down the costs of transportation, gas, liability, etc.
It might also have an impact on health care costs, as accidents cost the state, insurance, and patients, lots of money in hospital care for accidents. Certainly avoiding the negative economic impact of losing valuable people (aren't we all valuable?) to car crashes will also probably fuel legislation that makes it ever harder/costlier/illegal for a human to drive a car on a public road.
After all, replacement children don't come cheap!
http://youtu.be/GYSfncB4peU?t=1m25s
.
Prisencolinensinainciusol. Ol Rait!
Traditionally if a professional engineer signs off on something they're putting their personal reputation on the line.
I'd expect most hardware components in a car would be signed off by a PE. Software probably not so much since we really don't have a robust way of developing software yet.
I've seen some of the procedures for the space shuttle software. It's probably some of the most bug-free software on the planet, but the process required to change a line of code is onerous. There's no way you could put out a new car model every year while following those procedures.
Have the car give increasingly strident warnings if maintenance isn't done, finally shutting itself down. (Or at a bare minimum, have it broadcast a signal saying maintenance hasn't been done, and maybe it wouldn't be allowed on major roads or something.)
A self-driving car is going to have a huge amount of telemetry. It will be able to tell you more about the accident than any participant or observer. It will likely have video of any event (cache), plus all of the IMU data. Not only will it be able to show you that the light was green when it proceeded through the intersection, it will be able to calculate the speed and angle of the car that hit you once the vehicle type is plugged into the vector equation.
And you're right - if the system is found to be faulty the insurance company will have their lawyers go after the manufacturer. And the manufacturer will probably lose, or will settle to cover up the issue. EULAs mean nothing against gross negligence, and a good lawyer and stable of expert witnesses can always find gross negligence if there has been actual wrongdoing.
Is it just my observation, or are there way too many stupid people in the world?
Considering they are 'people' according to law. So they (the companies) are driving the car....
FYI, as for the guns don't kill, people do argument, gun companies are, afterall, people too. Just saying...
Because of the legalities, the automated cars will drive like grandmas.
Many of the kids today are MUCH more stupid. Go back just 50 to 70 years and kids as stupid as those today got eaten by wolves, killed by disease, or any number of natural Darwinian type deaths. Today we preserve the seeds of our own destruction by ensuring terminally stupid children survive until they can mate with ANOTHER surviving terminally stupid child and produce STUPID SQUARED kids. The movie Idiocracy
http://www.imdb.com/title/tt0387808/ was right on the money.
errr....umm...*whooosh* *whoosh* Is this thing on ?
If ALL cars were automated there should be no car to car accidents as they will all communicate with each other and at the first sign of trouble by one car all the others in the vicinity will know to slow down. But, in a hybrid world where the auto cars have to deal with humans we will have a lot of problems. 1. The auto car will be programmed to be nervous nellies. Since they cannot predict what the crazy humans do, they will stomp on the brakes at any sign of trouble. 1.1 I expect Humans to be very confused by the behaviour of the auto cars and cause accidents. Who gets blamed then? 2. I expect Humans to maliciously mess with the auto cars, brake check them, etc. 3. Unless a unified auto car communication and behaviour protocol is created I expect many of the auto cars from different manufacturers to exhibit behaviours that will look bizarre to a human. 4-way stop sign resolution for example.
The problem of liability is relatively easy to solve. Define a set of operational standards, then limit liability by statute as long as those standards are met. Update and refine the standards as we gain experience.
A few unfortunate people will be killed by programming errors and deficient standards, but far more lives will be saved by getting the deadly menace of human drivers off the road.
A front seat driver of course!
Greed is the root of all evil.
today? no.
eventually? yes.
“Common sense is not so common.” — Voltaire
Keep waiting. For all the usual reasons.
We know where leadership by an anti-intellectual "strongman" who scapegoats minorities and likes boisterous rallies goes
That's a valid perspective for civil liability. Now what about criminal liability? Do I go to jail for manslaughter if the computer miscalculates and accidentally drives over a pedestrian? That's the problem.
Sure I'm paranoid, but am I paranoid enough?
So basically my fear with self driving cars is that they'll be like a friend that seems to always seems to stop late and be on the bumper of the car in front of you, but never gets into an accident. You know the type where your grabbing the door or ceiling handle and cringing... And this is coming from someone that gets regular evil eyes when I get onto a crowed subway.
> what's the point of a self-driving car if you can't relax or do something else while 'driving?'"
It's early days yet, give it time.
There was no point in cars AT ALL when they first were introduced: they were slower than horses, you needed to bring a mechanic along with you 'cos they broke down every mile. Some cities you had to employ a guy to walk ahead of the car with flags to warn people. Utterly without practical use, they were.
They got better, laws adapted. You have to start somewhere though!
Perfectly Normal Industries
Typ, that is already done in Europe - left lane for overpassing (for faster cars essentially), right line to be occupied if there is no need to use left lane. I have always wondered why American car users need to have three lanes each way none if which are prioritized.
So, no need for autonomous cars. Just some legislature changes.
Plain old sigh.
Sometimes driverless trains derail. Somebody's insurance company pays: either the manufacturers or the rail owners or the operators.
In the case of a car it probably makes sense to have the owner of the car buy the insurance, because this will be a lot lower than the insurance for non-driverless cars, because by the time they go mainstream, they won't crash as much as a person, because they don't get fatigued, drunk or thrill-seeky.
I've read many of the comments, and not one mentioned software failure, sensor failure, GPS failure, power failure, design failure, or hardware failure of any sort I can't envision. Or being hit by a non-automated car - which could make the system fail and drive the car into further danger. Is it possible techies cannot conceive of a computer system that does not work 100% of the time? (Makes me reflect back on all those posts I used to make contending that voting systems were inherently designed for cheating. No imagination. Machines *always work* in techies' view, it seems) A little too much programming - no experience in actual machines operating in the real world.
Computers *fail* in the real world. The more complex the system, the more certain the failure. An airplane can get away with automated flight, as there is room to maneuver and pilots are always standing close by. Cars have no safety margins for failure in traffic. None. This will not work, not unless people are willfully blind when the failures accumulate - possible.
What happens if someone spoofs a GPS signal? It's been done to drones, making them dive and kill themselves. What if a HERF gun blows out the brains of the car with EMP - or someone simply makes an EMP "bomb" and detonates it on an overpass?
Question again: who's responsible when the perfect machine fails and causes an accident?
THIS is why we shouldn't mix ethanol with driver fuel. Stick with straight cheeseburgers, I say.
I've fallen off your lawn, and I can't get up.
Exactly. If an insurance company is willing to stand behind the vehicle's operation, then any potential accident victims will be compensated.
Self-driving cars may even be a better bet for the insurance companies than selling policies for human-operated cars.
Skytrain may be on a separate track (mostly two, one for each direction.)
But you still need very careful traffic management to ensure that any train does not run into another one stopped at a station or anywhere else. I.e. a planned (station) or unplanned (random) stop.
These things go fast and hitting a stopped train would generate headlines. And if you where around Vancouver in the mid eighties when Skytrain was first built there where numerous naysayers predicting catastrophe if there was not an operator in each train to be able to "hit the brakes" if needed.
Back in 1860 nobody had even conceived of an automobile or the telephone.
They had plenty of laws for cutting telegraph lines, which did exist at the time.
When the telephone and automobile were each widely deployed, laws were created to address the social affects brought on by these advances in technology such as negligent vehicle operation and wiretapping.
The fact that the laws Had to evolve was a result of society recognizing the utility these inventions brought to bear.
The law is not a static thing, it is constantly evolving to match the needs and whims of society.
Insurance companies, on the other hand will require standardized testing at some point for policies to be granted. I suspect that the car makers have been developing such standards with the insurance companies, as such the law will have to catch up.
My prediction is:
Ultimately, autonomous cars will be subject to two standards:
1) if the on board computer/data logger indicates there was user intervention, and
2) if there was no user intervention.
No you don't. That is a wonderful goal to strive for, but we may come to learn that our AI just isn't gonna be good enough to have a totally driverless vehicle. That is a limitation of the system and it falls clearly into the deal-with-it category instead of the ban-it-cause-Im-scared-of-it category. You are making the mistake of thinking this is black and white like Troy up there is doing. That is flawed logic.
You are in control of your self-driving car. You tell it to go on the road. It isn't an act of god. It is a machine. It doesn't have a mind of its own. You (the owner) are liable for the damage it does. You can buy insurance to protect yourself from this liability.
The driver being liable does not imply that he has to stay alert or in control of the vehicle.
At best it implies that the driver has incentives to picki a safe vehicle, negotiate terms of insurance and pick manufacturers who offer guarantees.
Consider for instance the autopilot on airplanes or on elevators. Does the owner of the elevator have to stay and watch the elevator while keeping his hand on the stop button?
My guess is that in some cases the insurance company would pay and in some cases they would turn around and go after the car manufacturer.
Of course, most drivers don't have the expertise to evaluate the quality and safety of vehicles they wish to buy. This evaluation can be delegated to specialists (reputable middlemen and expert reviewers). Also if some autonomous cars have better track record, insurance companies could charge lower premiums on those and thus guide consumers towards safer vehicles.
These comments are mine; I do not speak for my employer.
Wrong. If the car is self driving the I'm not the driver. It's like if I take a cab or a bus.
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
There wouldn't be any criminal liability. The casualty would be classified as an industrial accident. The new sub department of the NTSB that handles automated cars would investigate the incident. The results would serve to change the protocols and make the self driving car industry safer. _IF_ an NTSB investigation shows gross negligence on behalf of the car manufacturer, that would open up the possibility for criminal charges against executives at the company.
As a side note. Do you think you would get criminally charged _today_ if you "accidentally" veered onto a side walk and ran over a couple of children?
Easy, the idiot that thought a car could drive itself, namely the driver, or lack thereof. Cars are meant to be driven, stick shift, fun. Not for taking a nap while on the way to wherever. Or perhaps playing a little backseat boogie at 70 mph with a lover.
Is the airline pilot still responsible for the aircraft when it's put on autopilot? Who is liable when a firearm of theirs discharges?
There wouldn't be any criminal liability.
I hadn't heard where there wouldn't be criminal liability if using an automated car. It makes sense if you aren't required to be alert and monitoring the car's performance while using it, but I can see it going the other way if you are.
Do you think you would get criminally charged _today_ if you "accidentally" veered onto a side walk and ran over a couple of children?
No, but if I was looking at my cell phone and hit a few kids in the street, I would. If I'm required to be alert while the automated car is driving, how would this be different? See the problem? I won't sign up to have to prove a computer glitch caused an accident or risk going to jail, but I'm more than happy to have a computer do the driving, far better than the average driver and better than most, if I don't face the risk of criminal liability if it does fail. Civil liability isn't a problem - insurance already covers that.
Sure I'm paranoid, but am I paranoid enough?
The interesting question is: What do cars think of the Trolley Problem?
It's not when software or hardware fails to do as intended that is the big problem when assigning blame. Nor when the car does something because it's been told to. It's the usually hidden and unused ethics which makes this difficult and interesting. Rules and duty versus utility and consequences.
It's most likely wrong to think that all self driving cars will act the same when faced with a necessary choice of who to kill and who to save. Unless governments require all such software to be written to very exact specifications about how to deal even with situations nobody has thought of, of course.
Should the cars' ethical system reflect their owners' views on the trolley problem?
Should cars be forced to use a specific ethical system contrary to what some owners would like?
Should cars learn by experience and adapt their ethics?
If you ride in your self driving car when it is, would you like it to kill you to save five others or the other way round?
Curious: CAN self driving cars see bicycles? If they can, do they know how to anticipate a bicycle's movements? How about recumbent trikes? Pedal powered vehicles of any time? Anything that isn't a car - how do they fit into the world of computing machines driving speeding tanks?
We are nowhere near ready for robot cars. Humans are general purpose computing machines that can perform pattern recognition tasks that no software can. If we are really concerned about human error to the extent we want to eliminate humans, then we should go back to formula and start building rail lines again. Making cars into trains is inefficient, not to mention impossible. A waste of time and resources in a world rapidly running out of both.
In the voice of James Earl Jones ...
In a world where Google cars could remove the death and danger of traffic in society, should we seriously consider the interests of lawyers who want to make us concerned about changes that could make the world a significantly safer place?
For fucks sake just work it out. Human driven traffic kills people. Google cars don't. Set up some insurance or something.
Sure it is great for MPG as there is less air resistance, but your front bumper paint will have to be touched up weekly!