What is the future for insurance when machines do what we do?

By Rick Huckstep

There is a series on British and American TV at the moment called Humans. The premise of the show is that mankind has developed the capability to build “Synths”, which are “human lookalike” robot machines that can reason and react and make judgments. These machines are operated and controlled by highly complex programs that run inside them whilst also connecting in some Internet of Things type way with everything around them.

The Synths are designed to perform mundane and routine tasks for their owners and masters. They are modern day servants that cook, clean, take the kids to school, care for the elderly and do all of the things that we do to run the home and our work.

Whilst this is the stuff of science fiction, the show has made it “real” by setting it in current times, not some futuristic vision of the world as it might be in the 22nd century. And whilst these machines are programmed to protect and keep humans safe, accidents do happen.

In episode 3, the lead character Synth called Anita throws herself in front of an out of control car to save a child. The Synth is “injured” (it’s odd as I write this how some words just don’t seem right when used in the context of a robot!)…I should say…the robot is broken and both the driver and the child are shaken but safe.

Clearly the actions of the intervening machine have prevented a serious accident.

So I ask myself…what would a world look like, where we rely heavily on machines for our routine and mundane tasks, for the insurance industry? And whilst we seem like we are way off the world portrayed by Humans, I’m not so sure it is.

You only have to look at the fantastic work of Hugh Herr at MIT Media Lab’s Biomechatronics group shown here in the TED talk filmed in 2014 (watch it to the end, it’s awesome!). And the artificial intelligence in IBM’s Watson, made famous 5 years ago by winning the TV quiz show Jeopardy in a shoot out with two human contestants. Whilst this was a fantastic gimmick, the seriousness behind Watson is that it will change the way human beings make decisions in many areas of life! Or the numerous examples on YouTube of robot competitions that range from them walking up a flight of stairs, to opening a closed door to picking up a drill and fixing a screw into a wall.

These all make the development of robot grass cutters look highly rudimentary by comparison…but this is just the start of it. Cutting the lawn is as routine and mundane a task as you get in the home and already a machine can take care of the bulk of this task..

When we look at the technological development of the motor vehicle, it is already pretty standard for these 4-wheeled machines to take care of the routine and mundane tasks. Such as drive at a consistent speed with cruise control, or applying the brakes before the driver has sensed the slower traffic in front. More than just controlling the acceleration and the braking, vehicles can now help the driver go in the right direction, e.g., Mercedes-Benz Lane Keeping Assist.

Automatic parking is becoming routine, reversing perfectly into a space that the vehicle has assessed to be of the right dimensions. Which in itself is great news for insurance companies with around a fifth of all motor claims coming from reverse parking mishaps alone!

We know that the car manufactures are going even further when it comes to introducing accident-preventing technology and this should be good for the motor insurance industry. Far fewer accidents mean far fewer claims and payouts, which is good, right?

But what about taking the leap to completely driver-less, autonomous vehicles? The National Highway Traffic Safety Administration (NHTSA) in the United States, has proposed a formal classification system to denote the level of automation in a vehicle. Level 0 is where the driver completely controls the vehicle e.g., classic cars, through to Level 4 where the vehicle performs all safety critical functions. E.g., Google, who plan to launch a Level 4 classified vehicle by 2020 where the car is full controlled by using laser technology. This car will be driver-less and could drive completely unoccupied (except in Germany where it is the law that a competent driver must be able to take over control of the autonomous car)

These Level 4 vehicles are not expected to have the kind of mishaps that us mere humans suffer from. And because they are communicating with each other they can all react to any eventuality in unison. The variable that will always upset the hand held apple cart is the human. Keep them out of the way, and the driverless car will never have an accident (unless it is hit by some inanimate object like a branch or a falling rock or an asteroid).

Nobody yet knows what the impact will be in the future, but various reports suggest that the number of motor incidents could fall by as much as 90%. Still good for the insurance industry? Possible not, because as motor claims fall, then so will premiums. By as much as 75% has been suggested.

However, these guestimates assume that motor insurance continues to protect the driver from liability for motor accidents.

In the driver-less car (as defined by Level 4), the driver is not controlling the car, which means they cannot be liable, can they? How will insurers assess risk and price for motor insurance against a driver where the driver history or their age or gender has no relevance? The premium should be the same for an 18 year old with no driving experience as it will for a 40 year old with 20 year’s clean driving history (assuming the same car, of course).

It is most likely that the motor insurance industry will have to adapt its business model. Liability will shift from the driver (as there isn’t one in a driver-less car) to the manufacturer (against the peril of mechanical or software failure). Consumers are unlikely to be any closer to a lifetime guarantee against mechanical failure as they are today. But it’s the software warranty that poses the biggest challenge for motor manufacturers. Software failure can be catastrophic.

One of the biggest challenges for software is currency. Anyone running a five-year-old computer will know that with every upgrade it increasingly runs like a dog and the number of incompatibilities increases with every new release. When your laptop crashes because of a software fault, it’s inconvenient, but for your vehicle…well, you can just imagine the challenges for manufacturers to keep the controlling software up to date across all its global fleet!

So the question of who carries the insurance liability is a tricky one. Risk is better placed when the risk can be managed/mitigated/controlled the best. But, by definition, the driver in the driver-less car has no control over the car so it must sit with the manufacturer.

If the manufacturers now become liable for insurance, how will that work in a global market place? Today, the link between the insurer and the insured is clear and easy to establish by jurisdiction. But if this shifts to the manufacturer, that link is muddied by international boundaries and the different legal and regulatory environments from one country to the next.

And if manufacturers were to underwrite the performance of their vehicles themselves, they become de facto self-insured businesses. But this is unlikely in the long term. However, for now and for a considerable period of time there will be a transition as autonomous vehicles are gradually adopted until they become the majority (assuming that they will, one day). During this period, the insurance industry is likely to pivot.  Instead of writing $bns of personal motor cover for the individual, they shift to commercial underwriting of the manufacturer’s liability.

Of course, if mankind continues their love affair with the combustion engine and the driving experience, the most likely outcome is a world with all 5 levels of vehicle sharing the road together (we will still see Level 0 classic cars on the road alongside Google’s Level 4).

This will also pose a challenge for the regulators and legislators alike and no doubt require global cooperation and standards to ensure some degree of alignment around the world. In a world where manufacturers, not drivers, are liable for claims, the rule-makers will need to legislate for autonomous vehicle manufacturers to cater for local conditions. For example, an American vehicle is involved in an accident in Germany, or South Africa, or Kuwait or China. The claim and any legal proceedings will occur in the country of the incident, not the US.

Maybe we will see a world where autonomous vehicles are only covered by the manufacturer for designated countries but not everywhere in the world!

The good news is that whatever route this complicated issue takes for the insurance industry, we should all be safer. And whilst risk may transfer from the individual to the manufacturer of the machine, overall there should be less of it (risk) in this world. That is the theory, at least, behind this vision of the future!


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.