Driverless cars – 10 unanswered questions

Nick Blacknell, Zurich’s Motor VM and Engineering Manager, considers 10 questions brokers may be asking.

1. When will self-driving vehicles reach our roads?
We’re not far away. Many high-end vehicle manufacturers, such as Tesla and BMW, already have elements of self-driving technology, and over the next few years we’ll start to see this technology filter down to mainstream cars.

All the organisations that I’m in contact with are predicting that by 2021, we will be at the stage where vehicles will be able to self-drive, with a driver on hand to take over if needed.

2. Will they be fully driverless?
While the UK has announced that a trial of self-driving lorries will take place before the end of next year, it is unclear if and when vehicles would be allowed on our roads without a human at the wheel.

By 2025, we could reach the point where these vehicles are capable of being fully autonomous, and in theory nobody would be required to concentrate on what the vehicle was doing.

3. Will autonomous vehicles be able to detect every possible hazard?
From potholes to snow to animals – not to mention other road users – it’s hard to comprehend the range of potential hazards autonomous vehicles will have to contend with. However, manufacturers of driverless vehicles are attempting to do just that.

Waymo, the successor to Google’s self-driving project, has tested some 20,000 scenarios on a Californian airbase and I’m absolutely sure we will get to the stage where autonomous vehicles will be able to detect every hazard a human driver would.

They will still face challenging scenarios, such as when a pedestrian steps into their path and the only way to avoid a collision is to swerve into a tree, or another vehicle. However, these are the same split-second dilemmas human drivers face every day.

4. What responsibilities will human drivers be left with?
The Automated and Electric Vehicles Bill, currently going through Parliament, will require drivers to install any essential software updates provided by the manufacturer. This isn’t likely to be a particular onerous task, however. The main responsibility drivers will be left with is taking control of the vehicle if it is unable to respond to an unanticipated scenario.

5. Who would be liable for an accident involving a driverless vehicle?
Human drivers could be liable if an accident occurs as a result of their failure to carry out the duties outlined above. However, if it is the result of software failure, then insurers could seek to recover costs from the manufacturer, in the same way they might today if a collision occurs due to a mechanical defect. When it comes to liability, things aren’t likely to change significantly until we reach a point where the vehicle is fully autonomous and making all decisions without driver override or intervention.

6. Could hackers take over an autonomous vehicle?
Chinese security researchers have already demonstrated their ability to hack a Tesla Model X and operate its brakes and doors. While those specific vulnerabilities have since been fixed, manufacturers face an almost impossible battle to make their software impenetrable to cyber criminals.

As soon as you connect a vehicle to the internet, you run the risk of it being hacked. That’s one of the downsides of connected vehicles.

7. What will driverless cars mean for fleet management?
The benefits could be wide-ranging, from route optimisation to reduced congestion and reduced fuel costs.

If we ever get to the stage where vehicles can drive completely autonomously, we could have a scenario where one car could be shared among multiple employees, driving them to their destination and then returning to a central base. There could be significant advantages for fleets there.

8. Will collisions involving vehicles really be a thing of the past?
Incident rates are declining already due to the advanced driver assistance systems we have in cars today. Some experts have claimed Autonomous Emergency Braking systems, for example, could save 1,100 lives in the UK over the next ten years.

We saw a similar impact on accident rates when traction control and ABS brakes were first introduced on cars in the 80s and 90s. I have no doubt that driverless technology will evolve to the point where the vast majority of hazards can be detected, leading to far fewer accidents.

9. Will insurers still want to insure non-autonomous vehicles?
Yes, because even when we reach the point where vehicles are capable of being fully autonomous, we will still have large numbers of human-driven vehicles on our roads. It’s inevitable, however, that the cost of insuring those vehicles will increase, as they will represent a greater risk than driverless vehicles.

10. What other impacts could driverless cars have?
For society, one of the biggest changes could be a sharp decline in vehicle ownership.

For the insurance industry, one of the biggest challenges we will face is the cost of repairing self-driving vehicles. It already costs a lot to repair vehicles with driver assistance systems, because they have a lot of expensive technology packed into the front end – the part of the car that is usually damaged in a collision.

As cars become more and more reliant on complex technology, insurers will have to manage their supply chains carefully, to make sure they are partnering with the right repairers with sufficient levels of expertise.