Telematic car insurance offers business opportunities to a new generation, but some fear a slippery slope toward big data discrimination
“We see you haven’t been in a wreck in five years,” says a car insurance salesman to a young woman looking to switch providers.“With our safe driving app, you could get a 5% discount on your plan,” he continues.
She, the client, accepts.
Equipped with a smartphone app that acts as a telematic device to track behavior when the car is in motion, the young woman goes about her next month, driving home from her job at a bar in downtown Chicago five nights a week. When she gets her first car insurance bill, there’s no 5% discount even though she’s been driving safely. She calls the salesman to inquire.
“Well, the system deducts points depending on the time you drive. A high percentage of wrecks happen between 1am and 5am after people leave the bars, so you lose points driving at that time.”
It seems the system puts some people at a disadvantage even when they’re not doing anything wrong, per se.
Behavior-monitoring, or black box technology, was the brainchild of the automobile industry with providers sensing a movement towards usage-based payments instead of monthly premiums. Whereas traditional insurance rewards ‘safe’ drivers based on their history on the road, usage-based insurance (UBI or pay as you drive, PAYD) relies on present patterns of behavior to determine a rate. These patterns can include what vehicle is being driven measured against distance, time, place and behaviors like speed and hard stops.
Telematic car insurance proponents say the technology puts power into consumers’ hands.
At first, insurance providers merely offered discounts on regularly priced premiums for safe driving, which could incentivize drivers to embrace safer practices. Plus, insurance companies benefited commercially from the use of this behavior-monitoring, allowing them to better segment customers and charge by risk profile. As the industry has evolved, providers are moving away from using the data as an incentive and towards utilizing the sensors as a way to determine a customer’s core price, which could fluctuate monthly or even daily.
Nigel Walsh, a partner at Deloitte UK focused on InsurTech, describes this as the next evolution for the insurance industry, giving insurers a new role above and beyond the claims business to one of proactive risk manager. “Insurance providers can create something unique and specialized quite quickly, offering different policies to different demographics with different costs associated with them.”
In this way, the industry moves away from a service based solely on average risk, and consumers can choose policies not just based on price like they do today, but on the service they’re getting on top. It’s a paradigm shift where insurance isn’t what you’re buying but a byproduct of an overall monitoring service that maybe offers car breathalyzers, home-security systems and wearable bracelets that shock you if pick up a candy bar in the grocery store. Walsh, a self-proclaimed optimist, sees future relationships with insurers in the following terms: “I pay you to monitor my life and make it better.”
“I pay the insurance company monitor my life and make it better” – Nigel Walsh, Deloitte UK
But in the night-driving example, the rules don’t categorize people all that correctly. The rule lumps teenage babysitters and overnight nurses into the same category as possibly drunk-driving bar-goers. The babysitter would have to decide whether to take a hit on insurance or take an hour’s less pay from his or her client. “You just need to understand the rules of the game,” Walsh says. “It’s a specific cause-and-effect to stop the young person from driving late at night.”
But is it causation, or correlation? This is the point for others who are not so excited about the future’s algorithmic overlords. Correlation is mutual relationship between two or more things, while causation is the action of one thing causing another. Driving late at night doesn’t cause wrecks. But there is a correlation between late-night driving and wrecks, for a number of reasons, including alcohol-consumption, low visibility and tiredness.
Back to the future?
“Based on how data points are measured, insurance algorithms can incorrectly categorize certain behaviors and associate those with certain risks,” says Anthony Lewis, a former independent consultant and author of bitsonblocks.com. “But there’s a miscategorization of risk-based behaviors, and will therefore be a mispricing of certain things for certain people.”
“There’s a miscategorization of risk-based behaviors, and will therefore be a mispricing of certain things for certain people” – Anthony Lewis, author
While Lewis can see why the insurance industry might want to move in this direction, there could be a problem when data shows us something we don’t want to see. What if telematics reveal that a given ethnic group are worse drivers than any other demographic, for instance?
And while some discriminatory practices can be found to infringe on individuals’ statutory rights, insurers have previously found ways to get around such legal safeguards. For instance, in December 2012, the European Union established gender equality rules that made it illegal to charge men a higher car insurance premium than women based solely on their sex. Insurance providers then slipped through a loophole, basing higher premiums on more male-dominated careers, with civil engineers, who are mostly men, charged considerably more than, say, the largely female collective of dental nurses.
The above scenario is just one example of the business transformation that big data technologists and other IT professionals are shaping in all segments of our economies. However, technological development and digital transformation do not come without dilemmas based on values. IE University’s Bachelor in Information Systems Management offers an ideal hands-on learning environment in which to grapple with the issues arising at the intersection of technology, the business world and humanities.