- 11 September 2012
- Latest News
- 3 Comments
-
-
-
Tweet
Solving Driverless Car Liability
11 September 2012 Posted By anthonyp
As we’ve reported a while back legal issues and liability could be big issues regarding driverless cars. For some interesting thoughts on legislation you should check out Bryant W Smith’s blog post on who actually drives a driverless vehicle.
Here I’m going to quickly summarise a few scenarios on how to handle liability when a driverless car crashes and kills someone.
1. Socialised Catastrophic Insurance.
This idea is similar to vaccine compensation schemes in various countries. Vaccines are great for public health, save lives, reduce disease related injury and are generally safe and well tolerated. However, there is a tiny chance (say, 1 in a million) that receiving a vaccine could kill or permanently disable the recipient. To account for this outcome, injuries from vaccines – particularly childhood vaccines – can be compensated. Caps on the payout and thresholds for the degree of disability are features of these schemes, which are funded from vaccine revenues.
Driverless cars in 10 or 15 years could be analogous to the vaccine example in that riding in a driverless car could pose very little risk to the individual. However, at some point an accident will happen and someone will be killed or injured. To compensate victims or their families the fund would award money for damages. The fund would likely be supported by revenues from selling driverless cars or by taxpayers.
The upside of this system is that car makers and programmers don’t have to over-engineer their products to be highly expensive and completely flawless to the point of paranoia. The downside is the same as the upside: should car makers and programmers not have to engineer their products to be flawless because they won’t be liable, there is less incentive to make the perfect product.
2. Caveat emptor - Owner Liability.
Basically this scenario is ‘Buyer Beware’. Legislation could be organised so that car makers are not liable for driverless car behaviour. As a result whoever buys a driverless car is taking responsibility as the driver of the car – solely taking the risks and responsibilities upon themselves should an accident happen, with legislation requiring all driverless car owners to possess the appropriate insurance.
The upside again is that its great for car makers. The downside is that driverless car technology seems a bit pointless if you crash your car and get sued by whomever you’ve run over or crashed into because someone at Ford or Google didn’t debug their software properly.
3. Caveat venditor - Seller Liability.
A far more likely scenario than #2 scenario of buyer liability and is probably the default legal setting in most countries today. This scenario is that not programming or constructing a driverless car properly is a form of tort. That is if a driverless car crashes due to manufacturing or programming error the manufacturer is liable.
The upside is that it really encourages car makers to be extremely rigorous in making their driverless cars. The downside is that car makers can get their programming right in 99.9% of situations but could still lose a lot of money in the 0.1% of situations they didn’t (or couldn’t plan for).
4. Travel insurance
This scenario is somewhat of a cross between 1 and 2 but instead of the cost being carried by society or one individual driver the cost is spread out by everyone using driverless cars. Its probably a more likely scenario if no-one owns driverless cars but we treat them as ubiquitously available taxis. The idea is that travel insurance would be charged as part of the taxi fare. If an accident occurred and the occupant was killed or injured the insurance company would pay out. Presumably caps and thresholds would also apply. Presumably the insurance companies would launch an investigation, if the owner* or manufacturer of the driverless taxi was negligent then they would also have to reimburse the insurance companies.
The upside is that this option could still be combined or augmented with option #1 of socialising increasingly rare catastrophic accidents. It also incentivises car makers to make less risky cars and continually debug their software as safer cars will attract lower insurance premiums. The downside is that travel insurance may be complicated and difficult to implement whilst also making driverless taxis less economical.
*Logically, taxi owners would also insure their driverless vehicles and keep them in good condition to avoid negligence claims and ensure payouts.
Related posts:
3 Responses to “Solving Driverless Car Liability”
I find all these arguments that liability issues will prevent the adoption of driverless cars as patently ludicrous. It’s like arguing in 1900 that the electrification of homes and businesses won’t happen, because it’s just too damn expensive. It was too costly to NOT electrify homes, and it’s too costly to not adopt driverless cars eventually.
We spend much more now on liability issues than we would with a driverless car solution, due to higher accident rates from human-piloted cars. How that would do anything but aid the introduction of driverless cars, is beyond me.
Sure, there’s no doubt that how liability is handled, by judiciaries and legislatures, will change with the advent of driverless cars. And maybe that will delay things a bit.
Maybe there’s a better term for this, but this is a classic hater mentality. Someone proposes a novel idea, that requires changes to how things are done. Maybe they’re worth it, maybe they’re not. But the hater can’t get past saying “Hey, that won’t work right here, right now. Therefore, that’ll never work!” The falsity of his logic is obvious; he can’t imagine greater changes being made to let the idea work, so he just reverts to being a hater.
The funny thing is, you’re seeing a lot more of these haters from the auto press than the automotive companies themselves. And the automotive companies have at least as much to lose, given that a robo-taxi solution might dramatically cut the demand for cars.
@TheMechanicalJerk You’re so right – great comment. If I had to pick another word it would be ‘killjoy’
Forgot to comment on your actual article. Great writeup – I wasn’t aware of some of these models.
Owner liability doesn’t make sense, because by definition, the owner has little responsibility in an accident, except in say, cases of neglecting the car’s maintenance schedule. Even then, a car company could institute warnings, alarms, and blocks to reduce the incidence of accidents like that.
#1 and #3 seem pretty similar, just one has the government collecting the money, and #3 has the automakers doing it. I say the automakers should collect the money and be responsible for making the payouts, because they are the ones who actually affect how accident-prone their vehicles are. Makers with a leg up on safety could then charge cheaper prices. However, perhaps safety will be so high across the board that there won’t be appreciable price differences due to insurance built in.
With #1, car makers that put in extra effort to maximize safety get penalized, because they don’t get any extra reward from the government. Meanwhile, other car makers can skimp on safety mechanisms and sell their cars more cheaply, all else equal.
#4 seems like taking the annual insurance cost and pricing it on a per mile basis. My insurance currently comes out to an average $0.16/mile. With a huge increase in safety, there’d be a huge reduction in insurance costs. I don’t see how this figure could do anything but come down; the transit authorities have already said that for driverless cars to hit the road, they must be significantly safer than human driven cars.
Overall, #3 makes sense for robo-cars that are purchased by the end user, and #4 for robo-taxis. If insurance remained a significant cost, eg above 10% of total transit cost per mile, car companies would compete fiercely on safety, as it would be reflected in fare rates.