New data sources and the algorithms that make sense of them are important building blocks for many financial products in emerging markets. These algorithms also raise new risks related to discrimination and ethics, which are not always understood, even by those overseeing them. There is a need to better understand how we can monitor and address ethical issues with algorithms such as fairness and discrimination.

In this webinar hosted by IPA, Cathy O'Neil, New York Times bestselling author of Weapons of Math Destruction and CEO of O'Neil Risk Consulting & Algorithmic Auditing (ORCAA), will present an overview of her research on “weapons of math destruction”—black-box algorithms being used with little or no accountability in decisions that deeply affect people’s lives, including financial services. Cathy will also introduce the ethical matrix, a flexible approach to identifying and mitigating the harms that can arise from these algorithms.

Jacob Appel, ORCAA Chief Strategist, will then dive deeper into the issue of proxy variables—seemingly-neutral data that contain information about individuals that the modeler is not supposed to know. Proxy variables are the reason why simple gender-blind or color-blind approaches are insufficient today. We will discuss a few current technical and policy approaches to dealing with these issues.

To place these questions into the context of financial services, our two panelists, Paul Whelpton, Chief Information and Analytics Officer for JUMO, and Malavika Raghavan, an independent expert on data protection in consumer finance, will speak to how digital financial services can address ethical challenges while still increasing access to financial services in underserved populations.

This webinar provides an opportunity to engage with leading voices on ethical use of consumer data for financial services, and learn from their first-hand experiences from work across the globe. The presentations will be followed with a Q&A moderated by Rafe Mazer, Director of IPA’s Consumer Protection Research Initiative.