Who’s to blame when your self-driving car crashes?

Don't blame me, officer. I was just watching a movie

As exciting as self-driving cars are, we still don’t know a lot about how they will eventually integrate into society. The technology would work best if all cars were autonomous but it looks like the first will share the roads with human drivers. Will they have steering wheels to allow manual control? Some concepts have removed the steering wheel entirely but most manufacturers look like they’ll allow manual control. These questions need to be answered before the 2020s, when the cars will supposedly go mainstream.

Most of the excitement around the technology focuses on the capabilities of self-driving vehicles, but it’s important to consider the legal and regulatory issues. If the driver is to blame, who is the driver? If the car is seen as the driver, who is responsible for it? The owner or the manufacturer? It might be your car that crashed, but you didn’t write the software that crashed it. And what will all this mean for insurance?

Who is driving?

Establishing who was driving when an accident occurs is already something that’s been investigated. A Tesla with self-parking abilities recently crashed and the owner claimed the car started driving without the “Summon” feature being activated. But Tesla reviewed the car’s log and found claim that the driver did activate the mode. This is just for a fairly limited parking feature where the car moves when told to. True autonomous cars will make decisions in real-time and hit something. Almost all of the accidents Google’s cars have been involved in have been caused by human drivers but there has been at least one incident where they accepted “partial responsibility”.

So establishing who was really driving will become paramount if these cars will also allow manual driving. Perhaps someday an illegal market will emerge for people who can hack the vehicles to change logs and make it look like the car was to blame. An incentive for such a market could be if insurance firms end up having to offer separate insurance agreements for manual and autonomous driving. A big question for the industry is whether the manufacturers will work with insurance companies or offer the insurance themselves. Either way, business and insurance models will have to change if the cars go mainstream.

Scaling up and insurance

The biggest changes to legal and regulatory hurdles will occur when the technology moves from a few pilot cities to several countries. Will allow countries (or states in the US) allow autonomous vehicles? There’s a lot of talk about self-driving trucks delivering goods across Europe but a single country can put a spanner in the works if they make the vehicles illegal.

The insurance issues might become even more problematic as the technology spreads and demand increases. Right now we treat cars and gadgets very differently when it comes to insurance. Car insurance is usually mandatory but tech insurance is voluntary. It’s not clear which direction insurance will go when it comes to self-driving cars. If it’s voluntary, the pricing of insurance will vary wildly compared to traditional car insurance, which is kept stable because it’s mandatory but could potentially increase dramatically if self-driving cars do turn out to be much safer.

Image: Google

Drivers are expected to keep their cars safe to drive. In the same way most countries have MOT checks to keep cars running smoothly, governments will probably have to introduce something similar where the software and sensors are tested as well as the mechanical parts. This is another area where blame will have to be debated when it comes to liability. If the car is poorly serviced and then crashes, who is to blame? Surely the owner is to blame if the car isn’t serviced frequently, but what about when it is deemed road-safe but it turns out that poor servicing led to the crash? Your guess is as good as ours on that one.

Current legislation

Various manufacturers, institutions, and governments have pitched in with views on how all of this will play out. Recently, the 1968 Vienna Convention on Road Traffic was amended to allow autonomous driving with the caveat that manual control is always available and the AI can be switched off. That’s bad news for the car concepts with no steering wheel but all the vehicles already on the roads or being released soon give the driver manual control. The way the United Nations sees it, you’re responsible for the actions of your car.

That seems pretty clear cut but the legal side of things will still be confusing when people choose to sue others. The UN will only allow the cars if people can manually take responsibility for them, but things like our current Consumer Protection Act allow us to direct blame at companies for defective products. This is why the insurance models could change drastically. Instead of personal motor insurance policies, it could very well switch to a type of product liability insurance for the manufacturers.

Focusing on the driving itself, it’s interesting to note that the laws aren’t really ready for the vision some of the manufacturers are proposing. We’ve seen concept cars where the passengers are facing each other so some have their backs to the front windscreen. As mentioned earlier, some of these concepts don’t even have steering wheels. In the UK, the current laws dictate that drivers always have a duty of care even when technology assists driving. This obviously addresses recent advances that allow cars to automatically brake and avoid collisions. But if a self-driving car was released to the public in the UK today, you would be legally required to keep your eyes on the road at all times and be ready to instantly switch to manual should the need arise. The car might be able to drive itself but you’re the one with the duty of care for it.


Today it looks like the drivers will be responsible when autonomous cars crash, but it’s difficult to make laws concerning technologies that aren’t even available. The considerations made for cars that can park themselves aren’t the same as those required for truly autonomous vehicles that will drive around busy cities. Yes, the laws in many countries have started to reflect autonomous technologies, but they’re still behind where the likes of Google and Ford will want them. Most amendments to laws are still focusing on assistive technologies like this:

We expect legislation to change quickly as governments work with the manufacturers. All of the big names in the emerging industry see the 2020-30s as the time they will set their robot cars loose on the roads and that is really soon when it comes to such a risky technology. None of them will want to spend millions on a product that ends up being unfit for sale due to new laws. It’s in their interests to cooperate and make things happen quickly.

We think this will result in the manufacturers having to step up and take most responsibility for the cars. It would convince legislators that the manufacturers take safety seriously if they’re willing to take the financial burden. Also, it would probably the best news for us. If the manufacturers are responsible for accidents involving the artificial intelligence or sensors, then they’ll be working extra hard to minimise the likelihood of incidents. Give them the responsibility, with a separate insurance for manual driving, and we should all be a lot safer.

Main image: Google