Q&A with Robert Peterson

Driverless car technology is moving ahead quickly. Are insurers keeping pace?

You often hear insurers don’t expect significant penetration for about 30 years. I think that is probably incorrect. If these cars are as safe and convenient as they are represented to be, there will be a lot of incentives for people to purchase them and for policymakers to encourage that because we’re talking about saving substantial numbers of injuries and lives. My personal view is that the fleet will turn over a lot faster than it has in the past.

What are the main issues for insurers?

Much coverage will move from personal auto to commercial insurance products. As cars are controlled more and more by the algorithm and the computer that’s in the car, if there’s a collision that is caused by a defect in design, that responsibility is ultimately going to be a products liability responsibility that will land on the original equipment manufacturer or some of the others in the commercial chain.

Imagine a car that would typically be driven 10,000 miles a year. If that car travels 1,000 miles in manual control and 9,000 miles in self-driving mode, then there’s only going to be 1,000 miles for which the driver might be responsible in an accident. That would drop that liability premium dramatically. There’s the main challenge, particularly for companies with very large books of personal auto.

What challenges do cars with shared-driving capabilities present?

If the car drives itself 100% of the time, then when the car misbehaves, 100% of the responsibility is going to fall on the OEM. If you have a shared-driving experience, there are going to be periods when the driver is going to be responsible. If there’s any fault on the part of the driver, they’ll want insurance to cover it.

How does taking people out of the driver’s seat entirely change the calculation?

In the more distant future we’ll have vehicle-to-vehicle communication and vehicle-to-infrastructure communication. You may have dozens of cars communicating when something goes wrong, and it may be impossible to figure out who is responsible for what went wrong.

Will we have to worry about car hacking?

Right now there is very little reason to hack cars. There’s no profit in it. That may change. Some people may do it just to be malevolent. The developers of these cars are worried about it, and they are trying to harden them to hacking as best they can.