Chapter 88. Auto Insurance: When Data Science and the Business Model Intersect
Edward Vandenberg
The business model of automobile insurance (and therefore of most large insurance carriers) is changing drastically. Seemingly every month, an innovation is announced that at some point will have all of us chauffeured around in autonomous vehicles that never have an accident and are never exposed to the weather. But there are clouds forming. Extreme personalization of underwriting risk, and therefore the premium charged, may not be fair and ethical for everyone.
When an accident occurs, people and societies depend on the financial backing of insurance to defray the cost of damages and injuries (yes, accidents and weather-related damages will still happen). The way this works today (in a simplified view) is that insurance companies create policies with generic ratings for different types of drivers and vehicles. The underwriting and rating are designed to spread the risk. That means that lots of drivers will pay a little bit more to cover the few who have accidents. This seems appropriate when you never know which driver will get into an accident.
But the way the industry has been moving for the last 10 years or so, with AI, machine learning, and lots more data, is toward individual rates that predict the frequency and severity of accidents and the specific ...
Get 97 Things About Ethics Everyone in Data Science Should Know now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
