Description
Background
Non-communicable diseases (NCDs) account for over 70% of premature adult deaths globally, with many risk behaviors such as tobacco use, harmful alcohol consumption, poor diet, and physical inactivity beginning in adolescence. In Kenya, increasing smartphone penetration presents a unique opportunity to engage youth with digital health innovations that promote behavior change and health literacy.
Methods
AfyaAI is a proof-of-concept mobile platform that leverages Large Language Models (LLMs) to deliver personalized, conversational health education to Kenyan youth. Co-designed with local stakeholders and youth representatives, the platform integrates AI models (e.g., ChatGPT, Gemini) to enable real-time dialogue on NCD risk factors. A mixed-methods study was conducted across academic institutions, markets, churches, and recreational spaces, engaging 437 youth (aged 18–35) through structured surveys and focus group discussions.
Findings
Findings reveal high acceptability and usability of LLM-powered tools among young people, with participants demonstrating curiosity and trust in AI-generated responses. Significant differences were noted across age, gender, and location in terms of content accessed and perceived relevance. The interdisciplinary Community Advisory Board (CAB) played a crucial role in ensuring ethical deployment and contextual appropriateness.
Conclusion
AfyaAI shows strong potential to serve as a low-cost, scalable intervention for health education and behaviour change in LMICs. Insights from this study highlight the need for further research on how youth interact with AI systems for health, and how these tools can be optimised for equity, gender responsiveness, and long-term engagement.
| Country | Kenya |
|---|---|
| Organization | Private Sector |
