Abstract:
Mental health issues affect university students, particularly from developing countries, who are mostly faced with
social, economic, and environmental challenges, which significantly impact students’ academic performance and well
being. Furthermore, the literature shows that seeking professional mental health support is usually associated with
excessive costs and stigmatization. Moreover, existing generic Chatbots do not address the specific contextual needs
of students. This paper therefore presents a Chatbot tailored for students from a developing country context.
Qualitative data were collected from ten final-year undergraduate students with a mental breakdown who are currently
enrolled at a single university case in Zimbabwe. A thematic analysis of collected data provided key constructs for
designing a Chatbot that employs natural language understanding techniques to afford university students a convenient
platform for acquiring mental health support. Kanban methodology was employed to develop the machine learning
Chatbot using the Rasa open-source framework which is a development framework that provides open-source tools
and libraries for building conversational chatbots, leveraging a collection of different pre-trained and customizable
machine-learning models for varying tasks within the chatbot such as intent classification, entity recognition, and
natural language understanding. The findings from this research reveal that university students appreciate this
innovative approach that promotes a private and user-friendly environment for their mental health care. The findings
also suggest that students who experience mental breakdown only prefer to use Chatbots as an alternative intervention
rather than a replacement for mental health professionals due to the unpredictable performance of chatbots, which
sometimes results in inconsistent and varying solutions for mental health care. These findings add value to the
knowledge body and impact university students who can now freely access mental health support without
stigmatization. The parents or guardians of affected students can now be relieved from incurring high costs for some
avoidable professional mental health services for their children. Lecturers can lecture and impart knowledge to
mentally relieved, present, and attentive students. The findings also add to the university management’s understanding
of the student's mental health needs and, can be able to strategize policies aimed at minimizing students’ mental
breakdown and recurrences.
