AI chatbots and students’ mental health support: An efficacy review

Abstract

Mental health challenges affect every facet of student life. The potential of AI tools, such as chatbots in mental health support, is particularly compelling given their ability to overcome traditional barriers to care. The study aimed to determine the effectiveness of AI chatbots in supporting students’ mental health. The study employed a systematic review following the guidelines of PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses). Relevant studies were identified through a comprehensive literature search across EBSCO (ERIC), Web of Science, and JSTOR. The study revealed that AI chatbots have the potential to reduce stress through automation. They increased engagement, cognitive achievement, self-efficacy, learning autonomy, and decreased frustration among students. Additionally, leveraging advanced machine learning models like GPT-based architecture combined with emotional AI could enhance the accuracy of emotional assessments to improve students' learning outcomes. The "how" of chatbots in supporting students' mental health, as explored in the present study, provides valuable insights for stakeholders in education. It highlights the various aspects of mental health issues that chatbots can support and guides the development of necessary chatbot tools for educators. Furthermore, the study aims to inform government and educational stakeholders about how advanced AI chatbots can impact students’ emotions toward learning. The recommendation is that AI chatbots should be integrated into school platforms or applications already in use by students to maximize their impact on mental health.