Identifying the Smarter Risk

Statistically speaking, a 16-year-old boy is a menace on the road compared to, say, a 40-year-old woman.

Insurance companies’ rates are built from massive amounts of historical data to reflect the increased risk they are taking to cover those more likely to be responsible for an accident. But what if your 16-year-old son actually does drive more like a responsible 40-year-old woman? Is it fair for him (or you) to have to pay higher premiums?

Artificial intelligence (AI) is addressing this conundrum.

Insurance companies are adding AI and machine learning (ML) systems to work alongside human talent to help the industry more accurately predict risk and provide more personalized rates than ever before.

“It is fascinating what's happening with the insurance industry,” said Kristian Hammond, director of Northwestern Engineering's Master of Science in Artificial Intelligence (MSAI) program. “Not only is AI impacting it, but we can see the beginnings of a real transformation.”

And this isn’t just with car insurance, Hammond said. AI’s ability to analyze unprecedented amounts and types of data is changing how the insurance industry builds rates for everything from life insurance to health insurance and fire insurance.

Few industries rely more on data than insurance. Historically, professionals called actuaries used reams of data to manually create statistical charts based on different demographic information, such as age, gender, and ZIP code. These actuaries reviewed each individual insurance coverage application to determine which spot on the chart a person belonged.

A teen male in a high-traffic ZIP code presented more of a risk than an adult female living in the country and would pay a higher premium for coverage, for example.

But now, a combination of AI and mobile technology is allowing those groups to be more precisely defined.

For example, an Apple Watch could feed insurers health information such as how much the wearer exercises and how well she sleeps. This would allow insurers to go beyond mere age and gender to provide one 30-year-old woman who moves a lot and sleeps well with a better rate than a 30-year-old woman who rarely exercises and sleeps poorly.

“We are approaching a world in which the transformation is going to be that if you go for a walk for a half-hour a day, we will drop your life insurance rates and your medical insurance rates,” Hammond said. “We will know that you will be healthier, and we don't have to worry about you dying or getting sick as much as someone who doesn't exercise.”

This isn’t about punishing the couch potato or the more aggressive driver, Hammond said. Rather, it provides an incentive for people wishing to save money to adapt their behaviors to be a better risk for an insurer to take.

The pushback the industry is facing, Hammond said, comes from those who feel such information sharing is an invasion of privacy.

Data privacy is an ongoing conversation in MSAI. When it comes to insurance, students learn from faculty, industry speakers, and hands-on projects about how AI can facilitate more accurate risk decisions.

This conversation comes at a key time for insurance companies.

The US Bureau of Labor Statistics projects that the insurance industry could lose roughly 400,000 workers through retirements by 2026 in the United States alone. Many of those jobs will not necessarily be filled by traditional actuaries but either by AI experts or actuaries with AI proficiency.

“For us at MSAI, students need to be thinking about how that machine learning system is going to be situated in the business model,” Hammond said. “It's absolutely changing the business model of what constitutes insurance and transforming it into an advisory system where if you do this, you'll live longer and pay less and if you do that you'll be safer and pay less. That's thrilling.”