Author: Duncan Minty
Publisher: Scandinavian Insurance Quarterly (SIQ)
Publication Year: 2017
Summary: The following article discusses how the insurance industry has long been using historical data to underwrite, price, and manage risks. The advanced algorithms that insurance firms rely on to profile customers underwrite and price premiums are built upon such data, and recent studies revealed that they are biased and can produce discriminatory results when used to train models. That is why it is important for insurers to adopt a structured approach to ethical issues.