Leveraging AI Systems in Depression Treatment: Underlying Mechanisms, Design Characteristics, and Facilitating Factors

Description

Mental illnesses, such as depression, have become a global medical crisis. Social stigma and insufficient mental healthcare workers are two major barriers for depressed patients to receive effective treatment. Leveraging artificial intelligence (AI) -enableddepression treatment system (AI system) is a possible way to remove these barriers because the presence of an AI system (vs. physician) can alleviate patients’ perception of social stigma, and AI can interact with a large number of users simultaneously. However, patients’ decision to accept an AI system depends on many other factors. Therefore, we propose to adopt a patient-centered approach to investigate the independent and joint effects of social stigma, task uncertainty, AI characteristics, and facilitating factors onpatients’ decision to adopt and use the AI system. Study 1 will investigate patients’ preferences in care providers (i.e., AI, physician, and AI–physician hybrid) in depression treatment. AI–physician hybrid refers to an AI–physician task assembly in which the AI system interacts with patients at the front end and physicians verify the diagnosis results at the back end. We will also investigate the impacts of two contingent factors (social stigma and task uncertainty) and the possible underlying mechanisms (i.e., privacy concern and trust) that lead to the patients’preference in care providers. Based on the results of Study 1, we expect that the barriers (e.g., low trust) in the use of AI or AI–physician hybrid in depression treatment can be identified. Study 2’s overarching objectives are to examine how the barriers could be mitigated or resolvedthrough AI characteristics (empathy and explainability) and facilitating factors (company reputation and government regulation). Thus, Study 2 will explore the impacts of these AI characteristics and facilitating factors on patients’ initial use and reuse intention. Care providers’ empathy is a key characteristic that patients with depression value the most during consultations for depression treatment. Thus, we will investigate how AI empathy can alter patients’ privacy concerns and trust, which affect their initial use. Furthermore, we will explore how AI vendor reputation and government regulation can facilitate the impacts of empathy on privacy concerns and trust. Lastly, we will examine whether patients’ reuse intention will be contingent on AI explainability. Collectively, the results of these studies will benefit AI developers, public health policymakers, hospitals, and, most importantly, patients with depression. The outcomewill lead to higher general well-being, a healthier workforce, fewer healthcare costs, and higher gross domestic product (GDP).