Joanne Hinds

Computer-based personality predictions - how usable are they?

Over the last decade, predicting personality using digital data has become a popular and highly sought-after means to gain understanding of people's feelings, attitudes, preferences, and behaviours. As such, academic research has proliferated and organisations are increasingly seeking to reap the benefits of microtargeting information to different audiences. To date, research efforts have primarily focussed on developing algorithms to predict personality successfully. As a result, little focus is placed on developing robust and unbiased methods, which raises questions over their utility when deployed in practice. This talk will analyse the state-of-the-art in terms of computer-based personality predictions and will explore how effective they are in a number of real-life applications.