Supporting mental health self-care discovery through a chatbot

Abstract

Good mental health is imperative for one’s wellbeing. While clinical mental disorder treatments exist, self-care is an essential aspect of mental health. This paper explores the use and perceived trust of conversational agents, chatbots, in the context of crowdsourced self-care through a between-subjects study (N = 80). One group used a standalone system with a conventional web interface to discover self-care methods. The other group used the same system wrapped in a chatbot interface, facilitating utterances and turn-taking between the user and a chatbot. We identify the security and integrity of the systems as critical factors that affect users’ trust. The chatbot interface scored lower on both these factors, and we contemplate the potential underlying reasons for this. We complement the quantitative data with qualitative analysis and synthesize our findings to identify suggestions for using chatbots in mental health contexts.

Therapists’ attitudes towards the use of AI in therapeutic practice: considering the therapeutic alliance

“…Whilst such developments remain nascent, this is not something that will take over from current practices swiftly. There are, however, developments in this arena (Moilanen et al, 2023) that make use of machine learning and can complement existing practices.…”

Overview of Chatbots with special emphasis on artificial intelligence-enabled ChatGPT in medical science

“…It also implementing the self-care method discovery directly within the conversation. (Moilanen et al, 2023) 9. Chatbot Implemented as behavioral health during the COVID-19 pandemic; it is an algorithm-based, automated, and interactive artificial intelligence conversational tool that uses natural language understanding to employ users by presenting a series of questions with simple multiple-choice answers.…”

Chatbot Deteksi Awal Gangguan Kecemasan Menggunakan Dialogflow

“…Hal ini tercermin dalam tabel 2 yang mencerminkan persentase tanggapan positif dari para peserta penelitian. Pada riset sebelumnya mendapatkan tanggapan persetase negatif atau lebih rendah dari peserta penelitian [22]. Pada Penelitian ini mencapai tingkat akurasi deteksi awal gangguan kecemasan sebesar 84%, hal ini berbeda dengan hasil penelitian sebelumnya yang menunjukan akuras sebesar 72% [5].…”