If You're Not Driving, Then You Shouldn't be Liable for Accidents in Self-Driving Cars
Fewer and less destructive accidents also mean lower overall liability costs

Self-driving vehicles will be safer than human-driven vehicles (if they are not, people won't ride in them). Consider that data show that the the installation of Tesla's autopilot system makes its cars 40 percent less likely to crash than they were previously. So the company is now bundling cheaper liability insurance with its sales prices in some overseas markets. While the advent of self-driving cars will significantly reduce their number, accidents will nevertheless still occur. In such cases, who should be responsible for the damages incurred? Riders should not be since they would have no control over the vehicle. In fact, future self-driving vehicles will not have passenger-accessbile controls like steering wheels or brakes.
In his new law review article, "Automated Driving and Product Liability," University of South Carolina law professor Bryant Walker Smith foresees a "shift from a compensation regime for conventional driving that is largely premised on vehicular negligence to a compensation regime for automated driving that increasingly implicates product liability." Since the "driver" of the vehicle will be the software and hardware installed by the manufacturer, Smith argues that automakers will be held liable in the case of accidents. In fact, Volvo, Google, and Daimler have all already stated that they would accept liability if their technologies are at fault. In addition, self-driving cars will have fewer and less destructive accidents than human-driven cars currently do, so the overall liability costs will less than they are now.
Ultimately, Smith persuasively argues that liability issues are not likely to slow down the introduction of self-driving vehicles.
Editor's Note: As of February 29, 2024, commenting privileges on reason.com posts are limited to Reason Plus subscribers. Past commenters are grandfathered in for a temporary period. Subscribe here to preserve your ability to comment. Your Reason Plus subscription also gives you an ad-free version of reason.com, along with full access to the digital edition and archives of Reason magazine. We request that comments be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of reason.com or Reason Foundation. We reserve the right to delete any comment and ban commenters for any reason at any time. Comments may only be edited within 5 minutes of posting. Report abuses.
Please
to post comments
In such cases, who should be responsible for the damages incurred?
Oh, that's easy.
Bill Gates.
Shyeah, it's not 1998 any more... it's gonna be IOS or Android running most of these vehicles.
If someone puts a brick on a gas pedal and ties of the steering wheel with some rope, then jumps out of the car, and the car hits a pedestrian, who is liable? The brick or the rope?
Right. This article is asinine.
Or, lets say you have an autonomous ... guard dog, who gets out of the yard and kills a kid. Who's liable?
The responsible party is the owner of the autonomous whatever it is.
Yes. We don't think of it that way because when something like that happens they only sue the manufacturer because that is the deep pocket. But yes, the owner of a machine that harms someone is liable just as much as the manufacturer.
J&J: First, most autonomous cars will not be privately owned - most will be fleets providing transportation services. Second, you are correct if an individual owner screws around with the equipment and/or fails to maintain adequate maintenance including software updates, then they will share some of the liability.
Check out the economics of rush hours and coincidence factors before believing that fleet solutions will be satisfactory to customers.
exactly
First, most autonomous cars will not be privately owned - most will be fleets providing transportation services.
How do you know this? Why do you assume this?
People want to own their cars. This is why people in the US don't take the bus to the train station to commute. They get in the car that they own that's sitting in their driveway. They'd rather sit in bumper to bumper traffic than schlep to the bus stop or train station and wait on a bus or train. To go to the store, they don't call a cab or take the bus, they take their own expensive personal car (if they can afford a personal car, which most people can).
I mean, why isn't ZipCar taking off more in big cities? Sure it exists, but why isn't it growing by leaps and bounds? Because people would rather own their own car. It's a hassle and expense to keep a personal car sitting idle most of the time in your driveway or parking spot (that you might have to pay monthly for) but people choose that over having a ZipCar membership because they have control over their own car (or at least the appearance of control).
Word. If I could bring my own toilet to work, I would.
I think you're way off base on the 'most cars will be fleet-owned' idea.
That's simply not happening unless the costs of owning a private car are made so onerous that simple things like leaving the child-seat strapped in are no longer worth the expense.
For people who live in dense urban cores this can make sense - mainly because the cost of land is so high that its prohibitively expense for most people to use it for parking - but outside the cores, most people live in suburb and ex-burb enclaves. Some of us even in rural areas. A 10-15 minute wait for a vehicle for a trip that is already going to be 20-30 minutes long?
Or multiple trips. Let's say each intermediate stop adds an extra 5 minutes while waiting for the car to arrive - that adds up pretty quickly. Then what are you going to do with anything you've bought in between? Take it out of the car, carry it around with you, then load it back into the next car each time?
Car ownership is, right now, ridiculously cheap. Consider that people in most of the rest of the world use *motorcycles and scooters* as primary transportation while in the US they are mostly luxury items.
" First, most autonomous cars will not be privately owned - most will be fleets providing transportation services."
Unless these fleets are owned by the gubmint, they are privately owned.
Yes, Comrade! Forward into the future! Let us sacrifice some of our autonomy in the name of a little added safety.
History is littered with the bodies success stories of this sentiment!
Needless to say, I'm grateful that this technology won't come to pass in my lifetime no matter what the technocratic futurists think. A computer can't even identify a picture of a cat vs. a picture of a baby consistently let alone make value judgments on which one of those is ok to slam into at 70mph and which one is not.
Civil liability, but not criminal, right?
If someone puts a brick on a gas pedal and ties of the steering wheel with some rope, then jumps out of the car, and the car hits a pedestrian, who is liable? The brick or the rope?
That condition was created directly by the person placing the brick and tying the rope. This is not analogous to self-driving cars.
Sure it is.
Only if a google employee straps it in for you. Then it would be analogous.
You're right it's analogous. Assuming that you are considering the software programmer as the analogue of the person placing the brick and tying the rope.
If you want to kill self-driving cars as a concept, make the manufacturer liable for any and all accidents. Good luck with that.
Bingo. And that is how it is going to work and that, along with the network security issues, what is going to kill them.
They can always blame Russian hackers for a car crash.
Well, the article says manufacturers are all cool with paying up "if their technologies are at fault." It just means that their lawyers have already figured out how to never be at fault. Or, they figure that a couple million a year in payouts is an acceptable cost of doing business. I don't know.
As I explain below the "if they are at fault" qualifier renders the promise meaningless. Products liability, at least in the US, doesn't take into account fault. It is strict liability.
Oh, that will change. In any event, the will be comparative fault Berwyn multiple vehicles in a crash, or somewhat likely, it will be determined in many cases that no amount of foresight, prediction or automation could have prevented the accident.
One thing is for sure, the motor vehicle injury plaintiff's bar will have to get pretty technologically savvy, and the will be a cottage industry of technical experts for establishing faulty programming.
University of South Carolina law professor Bryant Walker Smith foresees a "shift from a compensation regime for conventional driving that is largely premised on vehicular negligence to a compensation regime for automated driving that increasingly implicates product liability."
That is true and likely how it is going to work. That is not, however, a good news story. Products liability is strict liability, even in cases of foreseeable misuse. You don't have to establish negligence to win a products' liability case. You only have to establish that using the product injured you.
Smith argues that automakers will be held liable in the case of accidents. In fact, Volvo, Google, and Daimler have all already stated that they would accept liability if their technologies are at fault.
Which, if you understand product liability law, is a meaningless promise.
Smith argues that automakers will be held liable in the case of accidents. In fact, Volvo, Google, and Daimler have all already stated that they would accept liability if their technologies are at fault.
Which, if you understand product liability law, is a meaningless promise.
I wondered about that myself. Volvo et. al. promising to accept liability if their technologies are at fault is going to involve a lot of lawyers claiming in court that their technology wasn't at fault.
Computers will be computers... accidents will happen. It is not the manufacturer's fault that the truck was painted white and the sun's angle made the truck blend in with the clouds.
As a functional analogy, consider things that are already automated and have been for some time: say, garage door openers, or elevators and escalators. An automatic garage door is not supposed to squish children... but they do, and when they do, there's a law suit, finger-of-blame pointing, etc. Ditto elevators and escalators: someone gets killed on one, the finger of blame points to the maintenance company, the building owner, the victim.
Self driving cars - even at one tenth the accident rate of human-driven ones - are likely to create a tort nightmare.
No way will the government like this cause He loss of revenue (red light, speeding etc fines). Or i suspect they wont think that far ahead...electric cars and cafe standards vs gas revenue
They will never drive well enough to make them attractive to most people. When you drive yourself, you determine how fast and what traffic rules to break and what risk to take.
Think of it this way, if I offered to pay an old lady who never exceeded the speed limit or so much as rolled a stop sign to come and drive for you, would you take the offer? Understand, she would drive you everywhere you needed to go but the price of that is you can no longer drive your car, even in an emergency. I think most people would pass on such an offer.
"I think most people would pass on such an offer."
When I go to work? I drive myself. The rest of the time, I'm happy to let whoever else is in the car drive, and do a fairly good job of ignoring their driving habits.
So yeah, I'd probably be fine with such an offer provided she wasn't chatty or smelled or anything like that. Throw in that it'd lower my insurance, improve my gas mileage, and brighten my smile? Yeah, that sounds like a good deal. Means I can spend more time on things I like to do and less time on things I have to do.
And here's the thing: I think you're being unjustifiably pessimistic. We have at least one US city where car ownership is at less then half the households (New York City), and having a car is decreasing as a status symbol among Millennials?.
And here's a bigger thing: if I was wrong, and people aren't happy handing off the wheel? Then we wouldn't have the slow progression towards auto-autonomy that we have. Cruise-control, stay-in-lane, reactive cruise-control, even that parallel-parking gimmick... that stuff sells for a reason.
________
?Twice a week I drive 90 minutes one-way in the morning and evening, for a combined six hours of commute over two days. The rest of the week I telecommute.
?This might change as more Millennials hit middle class and can buy fancy cars, but at least so far we're not as into cars as our parents.
"Understand, she would drive you everywhere you needed to go but the price of that is you can no longer drive your car, even in an emergency. I think most people would pass on such an offer."
That's why rich people don't ever have chauffeurs!
After walking away, I remembered the recent test drive of a highway-autonomous 18-wheeler in Colorado?. And that made me think of another reason people will be happy to let "granny" drive 'em around: no drunk driving if you're not driving. No more having to worry about designated drivers, or calling a cab and coming back for your car the next day, or having to judge "am I good to drive?" while your judgement is already impaired.
Face it, for folks that like to get shit-faced, self-driving cars will be a life-saver. Both financially and literally.
________
?Try google-searching "autonomous beer truck test". Should be a story from October 2016 about it.
I agree. It's the OTHER drivers who I worry about. I WANT them to drive like little old ladies. Better than driving like assholes.
You're assuming the gov will let the car operate with no manual backup.
Someone will be the DD, they just wont actually have to drive. Which brings us back to where we are now.
or call an Uber.
Fuck yeah. I wonder who will offer the first model with a built-in bar and fridge?
I'd pass on it for ME, but for YOU and everyone else, I think it's awesome.
Don't worry, they'll come up with other ways to make up the lost revenue. And they will when they realize how much control self driving cars gives them over people.
Self driving cars simply hands our freedom to go where we want over to the oligarchic silicon valley corporations and the government
At the very least, they will monitor where people go. At worst, they will limit it.
Yes and all that downside for mostly illusory benefits.
RE: If You're Not Driving, Then You Shouldn't be Liable for Accidents in Self-Driving Cars
If you continue to use logic, facts and common sense, then you will lose the trust of idiots everywhere.
You've been warned.
Anyone who is actually fucking stupid enough to believe that the vehicle manufacturers and/or parts manufacturers are going to be held legally liable in every case for the lifetime of the vehicle probably also believed that there was no way Hildog could ever possibly lose the election.
Better maintain absolutely perfect timing and paperwork records on all your safety inspections and part maintenance schedules. Because if you're even so much as one day late on your required inspection, and the sensor array on the front of the vehicle happens to pick that day to die on you and run over someone's six year old kid in the middle of the street, you will be one very sorry and fully liable self-driving vehicle. "Believe me", as the Donald likes to say.
Even if they were, the cost of that liability will be passed on to the consumer in the form of higher costs. So you will be paying for it one way or another.
I'm keen to hear what the oligarchy known as U.S. insurance companies has to say about this. I'd would have thought maybe Ron would looked into this for the article.
In fact, Volvo, Google, and Daimler have all already stated that they would accept liability if their technologies are at fault.
So you'll be battling two corporations with a vested interest in not paying claims.
What is funny is that Ron seems to believe that "if their technologies are at fault" isn't an exception that swallows the rule.
I think the insurance lobby will make sure that owners have to carry insurance one way or another.
It could be like homeowner's insurance. If a kid falls on my driveway and gets hurt, then they'll come after my insurance for the medical bills. It doesn't require negligence (like sloppy snow removal). It simply has to happen on my property.
This happened to my sister. A neighbor's kid came over to play with my niece and nephew. She fell and broke her arm and the neighbor's health insurance company successfully sued my sister's home insurance. There was no snow (it was summer) and the driveway was new & flat. My niece & nephew weren't even outside - the girl fell before she got to the door to ask them to play. There was no possible way to conclude negligence. Even so, the insurance companies went with "your land, your liability".
Why wouldn't insurance companies push for the same interpretation on self-driven cars?
And as Reason constantly reminds us, there is no such thing as a free lunch. One way or the other, you end up paying. Maybe the auto makers take on liability, but they build that into higher prices for the vehicles for consumers. That would be different than you or me paying out of pocket (or having our insurers pay out of pocket), but we'd still be paying in some fashion.
A lot of commentors are fixating on the liability part, and missing the overall point.
Even the cars we already have, that aren't fully autonomous, are enough of an improvement to reduce insurance.
So sure, companies will probably have a merry ol' dance dodging liability. But the cars will still be safer.
Yes. The whole point.
If you wish to remove liability from the owner/'operator', you're going to have to take away all control from them also.
Because *someone's* going to be liable - if not the owner, then the manufacturer. If the owner bears no responsibility for things like MAINTENANCE OF CRITICAL SYSTEMS then no one is going to 'sell' you car.
At best they'll 'license' one to you - in the vein of software - and you will not be allowed to modify it and will have to take it to seller authorized maintenance facilities for routine maintenance and repair. And on their schedule.
Failure to do the above will have to be a violation of the law. Something along the lines of the DCMA for autonomous vehicles specifically.
Product liability is shit law to start with. When negligent operation of a product can put the manufacturer on the hook - even if the operator is considered 99% at fault - its just a search for deep pockets to compensate people for what is mostly their own stupidity.
A: See my post above.
Won't that be cheap!
"First thing we do is kill all the lawyers"
As everyone has pointed out, the corporate weasel words about accepting liability actually mean nothing. There is first the point that this is about fully autonomous vehicles, no chance for the riders inside to influence the action(s) that resulted in harm. So they alone are in the clear. That leaves the vehicle, representing the software, the sensors, the actual mechanics (brakes, throttle, lights, tires, etc). The software could be at fault due to bad software design, poor coding, improper testing, poor update control etc. The sensors, and all the mechanical components are likewise subject to failure from multiple sources.
This will all come down to either a free market solution such as insurance companies defining the priority of faults, or government regulations defining a precedence of fault. With the current manufacturing and software industries both loaded with multiple outsourcing of components, this will be a liability lawyers gold mine until all of these factors are resolved.
And after 45 years in software design and development, I know way too much about the potential for failure.
(My solution if to stick with my high mileage, fully paid for, completely under my control, old Fords. Besides, none of the computer cars I have seen are convertibles.)
BMW 3 series - that's a sweet ride Charlie. Its all computerized nowadays *and* you can get it as a hard-top convertible (since 2008).
A more complicated question to answer is how liability attaches in a hybrid world of driverless, semi-driverless and human-driven cars? Human driven car and driverless car get in an accident: who's at fault? It will get ugly.
A more complicated question to answer is how liability attaches in a hybrid world of driverless, semi-driverless and human-driven cars? Human driven car and driverless car get in an accident: who's at fault? It will get ugly.
And right on cue, who's to blame for that double post? Certainly not user error!
We've been determining who is at fault in auto accidents since the first autos went on the roads, so how would a computer driver make it any different? Don't overthink this with product liability law; first you apply road law, and if the computer driver lost control or broke the law, the manufacture of that computer, software, sensors, and actuators is liable.
It has often been difficult to establish the facts, but now there will be a log of position, speed, and steering and throttle inputs in at least one car. Once the facts are clear, there are legal rules for liability. If a car caused a crash, the driver of that car almost always did something wrong, even if it's a computer - the log will show whether the computer or a human was in control, and that should settle the matter. If it failed to keep the car under control and following traffic laws, it didn't work right, and the details of that failure don't change the manufacturer's liability. (There might be exceptions for road hazards, missing or obscured signs and road markings, etc., where the government should be at fault for poor road maintenance - and GM or Ford will have a better chance of winning those cases than most people ever had.)