Author Identifier

Scott Hannah: https://orcid.org/0009-0005-6668-2926

Date of Award

2024

Document Type

Thesis

Publisher

Edith Cowan University

Degree Name

Bachelor of Psychology (Honours)

School

School of Arts and Humanities

First Supervisor

Joanne Dickson

Second Supervisor

Deirdre Drake

Abstract

As the mental health treatment gap widens and artificial intelligence (AI) gains global popularity, individuals are increasingly using the AI chatbot, ChatGPT, to seek conversational support for their personal mental health difficulties. This study investigated the use of ChatGPT for these purposes in an Australian context. A cross-sectional quasi-experimental group design investigated the association between self-reported mental health literacy (MHL), mental health stigma, and help-seeking intentions by adults who used ChatGPT for their mental health difficulties (User-Group) versus those who did not use ChatGPT for this purpose (Non-User Group). Mediation analyses investigated the potential associations between ChatGPT use, MHL, stigma, the perceived effectiveness of ChatGPT’s efficacy for supporting mental health and help-seeking intentions. A total of 455 undergraduate psychology students (n = 369) and community members (n = 86) participated in the study. After data screening, 397 participants were retained for analysis. Results found no significant group differences in MHL, anticipated mental health stigma, or help-seeking intentions between the User-Group versus the Non-User Group. However, ChatGPT users reported significantly higher levels of self-stigma than non-users. Mediation analysis results suggest that after controlling for age and gender, ChatGPT use positively predicts perceived effectiveness, which may, in turn, suppress anticipated stigma. The remaining results indicate that the unstructured use of AI chatbots offers no significant benefits in supporting adult users with their MHL, self-stigma or help-seeking intentions. Future research may benefit from investigating how AI chatbots can be optimised to enhance MHL, reduce stigma, and increase help-seeking in real-world applications.

Access Note

Access to this thesis is embargoed until 6th August 2026

Available for download on Thursday, August 06, 2026

Share

 
COinS