College students-in-the-loop for their mental health: a case of AI and humans working together to support well-being

Abstract

Technology plays a relevant role in mental health. Specifically, integrating pervasive technologies with artificial intelligence (AI) holds promising potential to collect users’ data, monitor individuals daily, and support treatment. However, the lack of trust in the collected data is a common limitation of prior work on mental health and technology. This paper proposes involving the user in a Human-in-the-loop approach as a solution to deal with the lack of accuracy of data collected through pervasive technology. In our study, end users judged and evaluated the data collected at two different times: before training the computational model, which would be later used for classification; and afterward to evaluate newly collected data that would be predicted and classified by the model. The solution proposed was implemented and tested in a project related to depression in college students. The results indicate positive reactions to the predicted classifications.