This was a qualitative assessment of a youth population in Kenya with engagement and attitudes towards NCDs and AI for health previously unassessed. Overall, we found an encouraging level of awareness of NCDs among youth in this population. However, this is hampered by deep-rooted stigma, spiritual beliefs, cultural perceptions, and economic challenges that govern behavior and delay access to care. There was a varied awareness of emerging AI technology with participants in urban areas exhibiting greater awareness of this technology and ways it could be applied to the prevention and management of NCDs.

Specific to NCDs, it was evident that spiritual and cultural interpretations, such as beliefs in witchcraft, curses, and evil spirits, continue to shape how NCDs are understood and managed, especially in rural areas. These beliefs often delay care and foster stigma, particularly around cancer and mental illness. Stigma manifests as social isolation, shame, and fear, deterring individuals from seeking timely care and contributing to late diagnoses and poor outcomes (11–15).

The adoption of lifestyle changes associated with prevention of NCDs is dismal. Despite increased awareness, participants reported that economic challenges, poverty, food insecurity, and daily labor demands make it difficult to adopt and sustain healthy behaviors. Many only changed their lifestyle after a diagnosis or health scare. Barriers to care included cost of care, distance to facilities, mistrust, and long queues. These issues are not uncommon in the region; additional studies in Kenya have shown similar patterns of high healthcare costs and access barriers. Similarly, in nearby Uganda, patients endorse identical concerns about unaffordable medical bills, transportation costs, long waiting times, and the absence of healthcare workers (16). Some solutions include enabling self-care and decentralized health care clinics, with successful models demonstrating feasibility of community-based integrated care for NCDs through differentiated service delivery approaches and task-sharing with non-physician clinicians (17). Similarly, in our study, AI and digital tools were viewed as promising, across all categories of participants, for their potential to support self-care through applications that promote medication adherence, symptom awareness, and health education, representing a novel and encouraging perspective in contemporary healthcare contexts (18)

Herbal and traditional medicine use remains pervasive, augmented by barriers to clinical facilities and beliefs about intrinsic efficacy of traditional remedies. However, this can cause delays in timely treatment, result in delayed diagnosis and contribute to poor outcomes for conditions requiring immediate biomedical intervention (19,20).

Participants identified CHPs as essential connectors between health systems and communities with a critical role in facilitating initial NCD care, as they have widespread access to a large number of community members; they provide health education, emotional support, and key initial screening and education. However, their successes in NCDs prevention and management is often constrained by a lack of authority, insufficient resources, and inadequate training(21). Further structural barriers on the part of the patient such as high costs, geographical distance, long queues, and mistrust in the healthcare system lead to failed referrals). Since CHPs are trusted and effective for initial engagement, policymakers should continue to ensure that CHPs have appropriate training, infrastructure, and resources to support and refer for chronic disease care.

. Although participants were optimistic about the potential of AI-powered apps for NCDs management through such as reminders and symptom checking, and the adoption of WhatsApp-based engagement platforms, digital inequality and mistrust in data use pose substantial challenges. Aspects of digital inequality that were highlighted include limited access to smartphones, high internet costs, and low digital literacy, which remain substantial barriers, especially in rural areas). This reflects previous findings showing that while digital health interventions are generally feasible and acceptable among youth in LMICs, persistent digital divides caused by cost, infrastructure, and digital literacy must be addressed to ensure equitable access (22–24). Importantly, participants emphasized that successful implementation of AI requires simple, localized, and culturally attuned tools that complement, but do not necessarily replace human care. Furthermore, in existing literature, digital health chatbots need to be adapted for the given context – including language options and familiar content; the AGILE chatbot for gender-based violence services in Kenya is one example that demonstrates the importance of user preferences and cultural adaptation in digital health tool development by adopting a human-centered approach in designing an effective chatbot with as many human features as possible to optimize utilization (25). In summary, AI-based tools provide a promising solution for health among Kenyan youth but will need to be adapted to the local context accounting for barriers in access. Advanced digital education can be employed to target youth in these contexts, including early in schools and at universities, to increase capacity for engagement (26).