No Safe Zone

Posted on April 21st, 2021 by Emilio Miles

In chapter 9 of Weapons of Math Destruction, Cathy O’Neil talks about how big data affects people’s ability to get insured. As has been discussed in the past, WMDs often target the poor and the less educated. She discusses how e-scores play a role in identifying these people and taking advantage of this vulnerability they possess. Many of them are desperate to drive — their jobs depend on it. In the eyes of insurance companies, overcharging them is good for the bottom line.

She talks a little bit about a statistician named Frederick Hoffman and draws comparisons between his work and modern WMDs. Hoffman worked for the Prudential Life Insurance Company. He published a report in the spring of 1896 that showed that African Americans had a shorter lifespan than that of other American citizens. This was used to make the case that black Americans were so precarious that the entire race was uninsurable. Modern insurance WMDs use the same concept. They group people who have similar driving behaviors into groups and charge them accordingly. These people are tracked on a map and evaluated on risk based on the area that they are in. For example, a person who drives by strip clubs and bars are more likely to gt into accidents since drunks are more likely to be out in the streets around these areas. Even worse, the driver himself/herself might be categorized as being drunk. Thus, you can see how insurance companies use a birds of a feather type approach towards pricing.

O’Neil also mentions how people fall into a negative feedback loop with insurance companies. Insurance companies tend to charge more money depending on a client’s credit score rather than their driving record. If an insurance company is able to overcharge a desperate and ignorant client, that person is more likely to default on other things like car loans, credit cards, and rent. This in turn lowers their credit score, which feeds back into the loop.

To make things more fair, insurance companies should stop looking towards behavioral data of a group. They should focus more on the individual. There are many times when putting people into groups can be harmful if a person in that group happens to be an outlier. On top of that, they should place less importance on credit score and more on an individual’s driving record. The biggest takeaway from this chapter is that even insurance companies make use of unfair WMDs and should be reformed.