AI Can Vanquish Bias: Algorithms We Can’t Understand Can Make Insurance Fairer

Author: Daniel Schreiber

Publisher: Lemonade

Publication Year: N/A

Summary: The following article discusses how the availability of big data and improvement in computing power now allows insurance companies to price more fairly in terms of premium per risk covered, but that does not mean they are not discriminatory. The author suggest using what is called a ‘uniform loss ratio’ test to check for bias. Simply put, if certain group has a lower loss ratio compared to the other groups, then the model may be biased.