Stakeholder Conceptualisation of Cyber-Risk and Priorities for the Implementation of AI Mental Health Chatbots
Year: 2026
Type: study-registration
Abstract: The use of AI chatbots for mental health and emotional support has grown rapidly in recent years, producing a increasing interest in academic literature to explore the ways in which people interact with this technology, choose to use this technology, and the harms associated with its use. Specifically, there is a growing body of technology adoption literature investigating attitudes and adoptive factors of mental health chatbots, however, a crucial aspect that has been underserved within academic and wider grey literature is cybersecurity. The cyber-risks posed by this emerging technology to user data, and how stakeholder perceive these risks is still poorly understood. The limited guidance produced by regulatory and authoritative bodies in cybersecurity and clinical safety compounds this issue by leaving a significant gap in knowledge at the intersection between cybersecurity risk, clinical safety, and user vulnerability. Previous research on this project revealed multiple prominent findings through a rapid review of the literature. Where user-level risk perception was studied in relation to technology adoption, mechanisms were inconsistently reported, displaying high end-user concern for data privacy risk that did not reliably translate to changes in behavioural action or chatbot usage. For user vulnerability, it is consistently reported that users reporting higher levels of mental distress or vulnerability are those more likely to engage with chatbots for mental health. This, compounded with our understanding that vulnerable users are less capable of detecting and responding to cybercrime, and often suffer most from its consequences, paints a concerning picture where it seems that those most in need of support are exposed to the greatest risk and consequence. This research is therefore part of a wider PhD project aiming to further our understanding of cybersecurity risk perceptions in the context of AI mental health chatbots in a way that could produce practical guidance for relevant stakeholders around the effective and appropriate deployment of this technology. In this study, we will aim to explore and understand the perspectives and priorities of relevant stakeholders on the implementation of AI chatbots for mental health support, probing specifically for their understanding of cybersecurity risk in this space, and any contextual factors that may influence perceived risk and how this translates to behavioural action. The following research questions will be explored: RQ1. What factors emerge as key stakeholder priorities for the appropriate implementation of AI chatbots in mental health support? RQ2. How do relevant stakeholders conceptualise/form mental models cybersecurity and cyber-risk in relation to AI mental health chatbots? Where do these views converge/diverge? RQ3. What contextual factors influence stakeholder perceptions of risk around the use of AI mental health chatbots? How do these factors impact their willingness to use or recommend the use of this technology?
Source: Open Science Framework
Authors: Thomas O'Fee, Dawn Branley-Bell
Language: English
Subfield: Applied Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7216373992
