Students’ perceptions of AI mental health chatbots: an exploratory qualitative study at Sultan Qaboos University

Comprehensive Summary

Here, AlMaskari et al. examined the perceptions of 60 university students at Sultan Qaboos University (SQU) in Oman towards the use of AI-based mental health chatbots to gauge any potential benefits as well as to understand the implications of their use. These perceptions were collected through a series of ten focus group discussions (FGDs) which were aimed at addressing the discrepancy in evidence for AI-based interventions in Oman. The FGDs paired with the authors’ use of thematic analysis provided three main areas of interest: benefits, challenges, and how to move forward. Among benefits, students listed the following: that AI chatbots are often more convenient and drastically increase accessibility to support, address the stigma surrounding traditional mental healthcare, have potential to monitor user activity to aid their treatment, promote self-led screening such that the user can determine if they need further help, and are more affordable than professional mental healthcare. Among challenges, students mentioned concerns related to privacy and data collection, lack of religious or cultural sensitivity, potential risks for dependency on chatbots, the potential for inaccuracies in the support provided, and the presence of bias in the data with which a chatbot is trained on when it comes to offering advice to users. Finally, the students offered how this technology should be taken forward; integrating a human in the interactions can ensure that high-risk cases as well as transitions to further care can occur without issue, maintaining past conversations and data from the user can aid in continuity of treatment, foundational ethical guidelines can ensure that users are not being taken advantage of, and wearable technology can provide a further dimension to personal mental healthcare by providing real time suggestions based on health monitoring.

Outcomes and Implications

AlMaskari et al. thoroughly explain the medical implications of AI chatbot use in mental healthcare through the testimonies of the university students who largely spoke of the beneficial implications these tools have in addressing the gaps in modern mental health treatment. The students’ suggestion for the involvement of a psychiatric professional in the AI-based chats addresses many of their initial concerns, such as lack of human connection and cultural/religious sensitivity, potential for inaccuracies, and presence of bias to name a few. The suggestions to further these tools by maintaining and using past data and to create wearable technology enhance their scope of use and applicability by allowing them to build a reserve of data that can make advice or suggestions more personalized and more available. And finally, the suggestion to establish an ethical framework and maintain transparency for the functioning of the AI-chatbot allows the user to utilize the resources with an understanding of where the advice is coming from without fear of companies using their personal data or taking advantage of them while they’re in a vulnerable state. Overall, the perceptions of these students were generally positive, and though there were plenty of concerns and limitations posed, it is established that with further development, AI-based chatbots could serve as a useful and valuable resource for mental healthcare.