Acceptability and implementation feasibility of a mobile application powered by generative artificial intelligence for mental health screening and referral in Uganda

Abstract Background Uganda faces a significant mental health burden, with depression and anxiety disorders affecting an estimated 14 million people. However, severe specialist shortages and high stigma create a significant care gap. Artificial Intelligence (AI) and mobile tools could offer scalable solutions to extend the reach of mental health services. Although AI-powered mental health applications are expanding globally, there is little evidence on their acceptability and implementation feasibility in low-income African settings. Objective This proof-of-concept study evaluated the acceptability and feasibility of a generative AI-powered mobile application designed for mental health screening, self-care, and referral in Uganda. Methods We conducted a mixed-methods proof-of-concept study between January and May 2026 in Kampala, Uganda, involving 50 participants (25 members of the general public and 25 healthcare workers). Participants evaluated a generative AI-powered mobile application integrating a large language model with validated mental health screening tools (PHQ-4, PHQ-9 and GAD-7) for screening, self-care and referral. Quantitative and qualitative data were collected through baseline and endline assessments. Acceptability was evaluated using the Technology Acceptance Model (TAM), while implementation feasibility was assessed using Bowen's feasibility framework. Results The application demonstrated high acceptability and implementation feasibility. More than 90% of participants reported being willing to continue using and recommend the application, while 95.6% found the information relevant to their needs. Healthcare workers viewed it as a valuable decision-support and task-sharing tool that could reduce workload and improve access to mental health services. The principal implementation barriers were internet dependence, mobile data costs, limited multilingual functionality, and smartphone access. Conclusion This proof-of-concept study demonstrates that a generative AI-powered mobile application is feasible and highly acceptable for mental health screening, self-care, and referral in Uganda. The application shows promise for supporting task-sharing, improving access to mental health services, and strengthening non-specialist care. However, successful scale-up will require addressing barriers related to internet connectivity, data costs, language accessibility, and smartphone access. Larger studies are needed to evaluate its effectiveness, cost-effectiveness, and integration into routine health systems in Uganda and similar low-resource settings.