The Potential of Conversational AI for Mental Health Support in UK Armed Forces Personnel and Veterans: A Mixed-Methods Health Needs Assessment and Feasibility Study

BACKGROUND: Mental health conditions represent a significant challenge for United Kingdom (UK) armed forces personnel and veterans. Despite comprehensive support services, persistent barriers contribute to a treatment gap. This paper explores the potential of conversational artificial intelligence (AI), or voicebots, as a novel intervention to bridge this gap. METHODS: A mixed-methods study was conducted to assess the feasibility and user acceptance of a voicebot intervention. The study comprised an online survey of 146 serving and former UK military personnel to gather quantitative data on technology use and comfort levels, and qualitative data on perceived barriers to care. This was supplemented by key informant consultations (N=4) with senior military and healthcare leaders to understand the institutional perspective. RESULTS: A disconnect was found between institutional perceptions and user-reported experience. While leadership often cited stigma as the primary barrier, users identified systemic and logistical issues, such as 'Time / Availability of Appointments' (60%), as the main obstacle, with stigma cited by only 8% of respondents. The mean comfort score (out of 5) for discussing health issues with a human professional was 4.0, compared to 2.6 for a voicebot. However, 54% of respondents indicated a willingness (scores of 3, 4, or 5) to use a voicebot for confidential health information, with acceptance paradoxically higher among older, more experienced personnel. CONCLUSION: A significant unmet mental health need persists within the UK military community, driven more by systemic barriers than cultural stigma. Conversational AI is a potentially acceptable adjunct to existing care pathways, particularly for an older, pragmatic cohort. Addressing logistical bottlenecks in the current system is as critical as developing new technological solutions. Rigorous evaluation of a voicebot MVP, integrated into existing frameworks, is warranted.