Will AI Kill The Need For Insurance?


Self-driving cars will eventually mean nobody gets killed on the road, which would be a stunningly wonderful accomplishment. So will smart devices kill our need for insurance?

After all, risk is the product of uncertainty, and insurance protects us against the risks of people making wrong or harmful decisions, things failing or malfunctioning, or the unreliability of nature (like trees succumbing to strong winds, or heart attacks).

An all-knowing artificial intelligence, or AI, could assess and then compensate for those uncertainties of experience, so no more car crashes or missed medical diagnoses. The myriad of life’s little nuisances could be prevented, too. The outcomes of every decision would be predictably reliable.

So why would anybody need insurance?

The answer is that we won’t, at least in some far-off future that is anything but, er, predictably reliable. In the meantime, it could be boom times for the insurance business, for at least four reasons:

The Learning Curve For Driverless Cars

Right now, the various autonomous vehicles on the road are the equivalent of 15 year-olds who’ve just got their learner’s permits. At best. There are going to be many years of mistakes before they’re safe, and the increasingly complex relationship between human and robot drivers is going to get dicier (“I thought you were in control, no, wait, what does that ding mean again?”).

Throw in the involvement of vehicle-to-infrastructure (“V2X”) communications, like cities trying to manage traffic flow, and the fact that some robot drivers may be dumber than others, and assessing and assigning responsibility (manufacturers, city managers, drivers, etc.) will provide the recipes for different risks.

Smarter Devices Mean More Complexity

Smart devices make decisions faster, which means they make more of them, and they also make many of them outside of human awareness (you don’t really want to witness the coding Alexa goes through to find your favorite music). So the efficiency of those actions might be tempered by their sheer volume: Imperfect decisions may happen too quickly to stop, or too quietly to review until some damage is done.

It might seem counterintuitive, but more smart cars on the roads, or devices in hospitals, businesses, and homes might very well mean more chances for those super-quick thinking robots to run into conflicts with one another. Those will be full employment opportunities for insurers.

Flipping Risk From A Cost To A Benefit

This one’s a little ‘out there,’ but imagine a world wherein driving is quasi-automated, yet some people will want the authority to drive. They’ll want risk, not want to avoid it, and insurers could charge them a premium like they do young men who own sports cars.

There could be others: People who don’t want to be monitored by healthcare wearables, or rely on the recommendations of investment bots for their retirement accounts. Home assistants shut out of making sure the hot water heater doesn’t explode. The prospect of a risk-free world might invent endless opportunities to sell heretofore unimagined risks as benefits to customers.

Insurers could invent entirely new products to monetize such desires.

What About AI Insurance?

Currently, an insurance policy is written and then updated once a year, assuming no event intervenes to change the terms. What if insurance coverage and rates changed based on moments of experience, and policyholders could manage their exposure to risk in real-time? What if policyholders had relationships with AI that managed their experiences, sort of like advisors or mentors? It could redefine what “life insurance” means by really insuring living.

The easy answer for the challenge of autonomy to insurance is to say the former will obviate the need for the latter. The truth couldn’t be more different, and far more exciting.

Too bad we don’t hear it from insurers.