AI Regulations - Testing Models for Discrimination Webcast (Bias credit)

Webcast Pricing

Our standard webcast pricing is the convenient way to access a single SOA webcast. Visit the webcast registration information page for details.

Laws against discriminatory insurance pricing have been around since well before Artificial Intelligence (AI) and Machine Learning (ML) models have gained traction. State laws and regulations define what characteristics insurers can discriminate on (age, gender), and which they can't (race, religion). However, with the availability of deep sources of personally identifiable data and the AI/ML modelling techniques to exploit them, regulators are concerned that disallowed discrimination may occur by proxy. The definition of what, exactly, counts as "discriminatory" and what proxies hit that are still being discussed.

This webcast will discuss the regulations, the underlying challenges with testing for discrimination, and some helpful quantitative techniques.

Achieve your annual Bias training with this webcast.