From Bias to Black Boxes - Understanding and Managing the Risks of AI
A Talk by Valerie du Preez FIA and Richard Galbraith FIA CERA
About this Talk
We examine AI risks through an actuarial lens, drawing on research from an industry focus group. Our goal is to identify challenges in integrating modern modeling techniques into traditional actuarial tasks while considering professional standards and regulations. We propose best practices and potential next steps to address these challenges, focusing initially on traditional actuarial perspectives and providing examples. Key themes include ethical considerations such as bias and fairness, professional concerns like interpretability, and the shortage of relevant skills.
