To succeed in an organization, people need to make accurate psychological inferences. Hiring managers must correctly identify the personality of potential recruits. Aspiring leaders ought to figure out how they come across. Executives have to recognize what others want and strive for. For most of human history, there was nothing else on the planet that could make these sorts of inferences as well as people could. However, rapid technological advances are enabling computers to understand the minds of others even better than we can.

In my research, I demonstrate that computers can not only form such psychological inferences (e.g., predicting what a new person is like, how others will perceive someone, what an individual wants, etc.), but that they can often do so with superhuman accuracy. Such developments hold tremendous promise for advancing both theory and practice in organizational behavior.

These technologies can advance theory by enabling more accurate models of human behavior than have ever been possible before. This is because, in order to understand a phenomenon better than humans can, a computer must have a theory that better corresponds with reality (i.e., is more accurate) than the one held by people. As I show in my dissertation, the greater accuracy afforded by these theories can illuminate relationships researchers have previously been unable to reliably detect. Moreover, we can use this heightened acuity to not only understand a particular theoretical relationship, but also why people are unable to perceive it. Thus, we can identify the psychological mechanisms preventing people from understanding each other (e.g., why people make mistakes when forming first impressions). 

In addition to advancing theory, we can also exploit these recent advances in computation to build tools that benefit individuals and organizations. For instance, in one project, I am exploring whether we can leverage this superhuman accuracy to help people more effectively navigate social interactions (e.g., by informing someone how she will be perceived during a first impression more accurately than she would have been able to predict by herself). In another, we built an online recommendation engine that automatically learns people's preferences and then provides them with customized recommendations for articles that help them become better versions of themselves.