Transparent algorithms in insurance

In this project, we developed methods for more transparent use of algorithms in the insurance industry, investigated the consequences of increased transparency, and studied ethical requirements on algorithms.

Insurance decisions come in many forms and can be based on different parameters. Different kinds of insurance also entail different preferences and different acceptance for which factors decisions may or may not be based on. In everything from life insurance for individuals to cyber insurance for companies, there are ethical dimensions. When humans are removed from the decision-making process, the need to understand how the algorithms work and make their decisions increases. Thus, automation of the right kind is a crucial strategic issue for the insurance industry.

Read some scientific publications:

• Towards Increased Transparency with Value Sensitive Design presented at the International Conference on Human-Computer Interaction
• Nordic lights? National AI policies for doing well by doing good in Journal of Cyber Policy
• Transparency and insurance professionals: a study of Swedish insurance practice attitudes and future development in The Geneva Papers on Risk and Insurance - Issues and Practice
• Explaining automated decision-making: a multinational study of the GDPR right to meaningful information in The Geneva Papers on Risk and Insurance - Issues and Practice
• Rawls’s Original Position and Algorithmic Fairness in Philosophy & Technology
• First- and Second-Level Bias in Automated Decision-making in Philosophy & Technology
• Transparent but incomprehensible: Investigating the relation between transparency, explanations, and usability in automated decision-making, PhD thesis from KTH.

Read the final report of the project here.