Talk delivered at Colorado State University on Aril 3, 2019
Slides available here.
These days it is not possible for even the best human chess player to beat the best AI chess player. But that hasn’t eliminated the game of chess. There are new leagues where humans play other humans while using AI companions to help inform decisions. Every chess move involves evaluating a large number of different options, which is something AI is very good at. Similarly managing an organization means balancing a huge quantity of complex information in order to make the right decision and achieve the right outcome. Managers frequently rely on intuition to make decisions in a timely manner. Machine learning models could augment that intuition by taking the thousands of organizational variables the manager must consider, reduce their complex interactions, and prescribe different courses of action. Just as AI has come to inform chess players, it will soon be a mainstay of managerial life. But what happens when a machine learning model tells a manager that one of their top employees is at high risk for quitting their job soon? What actions should she take? What are the privacy or ethical concerns?
In this talk I will discuss several projects I and my colleagues are working on using the network and contents of millions of emails from four different organizations to predict organizational attitudes and outcomes. On one hand we have the opportunity to answer broad questions about how human linguistic expression changes in response to organizational events or shifts in attitudes. On the other hand we must deal with the complex philosophical or ethical concerns of developing machines that could dramatically shift how our organizations are run. Just as the debate rages about the roles, capabilities, and consequences for self driving cars, I want the audience to consider the same sorts of questions for our companies, governments, and other organizations.