Who is liable? The driverless impact on insurance

Volvo's AutoPilot self-driving car

Who is liable? The driverless impact on insurance

Calculating risk and assigning liability with respect to today’s drivers, cars, and driving conditions is not a five-minute job. It’s a complex consideration dependent on many factors.

It’s even harder to calculate when it concerns the drivers, conditions, and autonomous vehicles of the future. But discuss and calculate it we must because it looks like the future is set to arrive in 2017, when Volvo launches 100 of its XC90s featuring their Autopilot function.

Volvo Autopilot driverless car

So how has Volvo tackled the question of who is liable when there is no human driver at the wheel? The Chief Executive has the answer:

Volvo will “accept full liability whenever one of its cars is in autonomous mode.”

That’s a bold statement, but it is perhaps something you might have expected to hear from the brand renowned for being the Kings of Safety. If they thought XC90 drivers would be worried about being held responsible if the car crashed while Autopilot was being used, Volvo has provided clear-cut assurance: they won’t be, as long as the Autopilot is at fault.

This is a radical move and sets a very clear precedent about where Volvo thinks the burden of liability should rest. But how have they come to this decision? The clue perhaps is in the name: driverless cars. For any insurer, the obvious consequence of their being no driver in a driverless car is whether individual drivers will still need insurance.

If the driver becomes a passenger, who will be liable in the event of a collision? The manufacturer? Which one? The company who designed, made parts for, assembled, or serviced the car? Potentially, any person or company involved in the process of creating the autonomous technology in a driverless car could in theory be held liable if that technology fails and causes a collision.

Volvo’s statement solves the question of liability, for owners of their XC90s at least. But this is just one car company making the first move. It’s not the UK government. Or the insurance industry. The jury is still largely out on the question of liability. But the issue needs urgent discussion because the answer could dramatically affect the insurance industry.

Factor in Google’s prediction that the adoption of driverless cars could bring about a 90% drop in the number of accidents and 90% drop in the number of cars on the road, and the consequences for the insurance industry could be profound indeed. Any change to the type of policies required, the specific wording of those policies, and the number sold could seriously affect the revenues and expenses of car insurers the world over.

How does the business community see things?

Management consultancy McKinsey & Company thinks insurers will shift liability from individuals to manufacturers during what it sees as the second phase of the adoption of autonomous vehicles in 2030.

Original equipment manufacturers (OEMs) will likely shoulder the burden of liability in future but as a consequence or requirement of doing so, they might also insist their driverless cars – or at least the autonomous technologies included within them – are only serviced by their own specialist teams.

Technology inside the self-driving Audi

OEM’s insurance premiums might also conceivably be lower than anticipated due to the risk-reducing, collision-preventing technology their cars would be using i.e. if autonomous vehicles (AVs) are safer and less likely to crash, then the risk is lower, thus the insurance premium should in theory be lower. Final payouts for claims might however be more expensive because of the cost of replacing the AVs’ highly advanced technology.

Lower risk, fewer collisions and lower premiums are all good news for the insured, but for the insurer – who might see their customer base shrink, and thus sales and revenues fall – such a shift in liability represents a risk to their existing business model.

How does the UK government see things?

To help a number of manufacturers carry out tests in 2015 on public roads, the UK government applied existing insurance law: the driverless cars must still have a driver who is capable of taking control and who must have insurance. Testing scenarios aside, the government is working with insurers, manufacturers, the EU and member states to figure out how best to insure autonomous vehicles and has outlined its thoughts and approach to regulation in its “Pathway to Driverless Cars” document.

Who is liable at which point?

The word driverless is in one sense a bit misleading. There will always be a driver, but it might not be human. It’s early days of course but it’s easy to see and agree with the idea that whenever humans aren’t in control of the vehicle, liability should switch to someone or something else: the manufacturer of the entire car and/or of its driverless technologies for example. Volvo’s statement and McKinsey’s analysis support this approach.

Driverless cars won’t suddenly arrive on our streets and replace all existing cars either. They’ll be introduced in stages. Over the coming years, we’ll most likely see an increase in the number of driverless features, like Volvo’s Autopilot, added to existing cars. Fully autonomous vehicles – such as the LUTZ Pathfinder or Meridian Shuttle – might appear at the same time as these incremental features but probably only in highly enclosed areas or zones like a motorway, pedestrianised street, a factory, or sporting arena for example.


In terms of liability, this could be a golden opportunity for insurers to develop new policies to meet the needs of hybrid cars and fully autonomous vehicles when and wherever they’re used. Existing law and policies could be adapted, rather than replaced wholesale. Policies will likely need to account for the fact that, during these developmental years, liability will shift back and forth between driver and a manufacturer multiple times during a journey.

What if…

The idea of shifting liability leads to another important question insurers need to consider: what if the driver chooses not to take back control when the car requests it? Or can’t? Who is liable then? Volvo’s XC90 will prompt the driver to take back control but if the driver doesn’t, the car will bring itself to a controlled stop in the lane as quickly and safely as possible.

All well and good you might say. But what if these interim features condition drivers to become less aware of a request or in some way less able to take back control? The more people become accustomed to decisions being made for them, the less likely they may be to make their own decisions. The design of the driving interface will therefore need to take account of habituation: a process whereby driverless cars might end up teaching us not to pay attention to the car and thus condition us to believe we don’t need to know how to control a vehicle, or be less able to when called upon to do so.

This effect might have a big implication for liability because even if someone is capable of retaking control but, through habituation, fails to, then can they still be deemed liable? Education of drivers about this risk and their responsibilities while the car is in driverless mode will therefore be critical. Manufacturers, aware of the risks of habituation, might also try to limit their liability by understating or limiting the car’s driverless capabilities.

Proving and disputing liability

Defining who is liable at which point is only part of the problem. Another question in need of discussion concerns what evidence will be needed to prove liability.

Whoever controls the data captured by the driverless technology in each vehicle, and in the wider driverless network as a whole, might hold the key to understanding the cause of any collision and assigning liability. One bonus of having access to such an array of data / evidence could be a massive reduction in the fraudulent claims. Fraudulent or not, if insurers aren’t able to access that data, they might not be able to determine liability as accurately as they could, meaning premiums might remain higher.

Key points for insurers

  • In the near future, liability will start to shift from individuals to manufacturers but not necessarily all at once or for an entire journey. At first, liability might only shift when a car’s driverless mode is activated and programmed by an existing insured driver.
  • In the time it takes to go from cars with some autonomous driving capabilities to fully autonomous cars, we will all still need personal car insurance of some description. Existing law, regulation and policies could possibly be adapted on an as-needed basis to cater for emerging driverless scenarios.
  • Once we reach a state where virtually all cars are fully autonomous, individuals might only need personal insurance against things like theft or lost belongings. Manufacturers of driverless technologies will likely bear liability while their products and services are being used and will need appropriate insurance policies in place.
  • This means the number and type of policies and customers per insurer will change, potentially leading to a loss in revenue for providers unable to adapt, and an increase for providers who are.
  • The effect per insurer will depend on their particular business model, customer base, product and service offering, insurance law and regulation, and the requirements of an evolving marketplace.

[separator type=”default-double”]

We may still be a few years away from fully driverless cars hitting our streets, but here at Adrian Flux we already offer policies for all sorts of driver-assistance technology, as well as practically every make and model of car, van, and caravan. Visit www.adrianflux.co.uk to find out how we could help you save on your insurance.