SaludIA: community health worker perceptions and implementation of AI-enabled integrated health-environment screening in rural Colombia

Background: Recent advances in Artificial Intelligence (AI) suggest that AI applications could transform healthcare delivery in resource-limited settings. However, little is known about community health workers’ (CHWs) knowledge, perceptions, and concerns regarding AI-enabled health applications when combined with environmental interventions. Objective: To examine CHWs’ knowledge, perceptions, understanding of AI, and anticipated benefits and challenges when integrating AI-enabled health-environment screening in rural Colombia. Methods: Empirical study conducted September-November 2025 with a purposively sampled cohort of 50 community health workers (54% female, mean age 46.8 years, mean experience 11.6 years) in Vereda El Manantial, Guadalajara de Buga, Colombia. Following a systematic two-condition video provocation protocol (positive scenario n=25, negative scenario n=25), participants engaged in semi-structured interviews averaging 45 minutes each, yielding a comprehensive qualitative dataset of 112 hours of audio-recorded material. Transcripts underwent dual-coded thematic analysis generating 156 initial codes synthesized through iterative consensus into 12 major themes. Stratified analysis examined response patterns across video conditions and demographic stratification by gender, age, CHW experience, and smartphone access. Results: Drawing on 112 hours of qualitative interview data from 50 CHWs, analysis reveals the Gender Paradox in technology adoption: male CHWs demonstrated skepticism-then-utopianism (65%→70%), while female CHWs maintained pragmatic gradualism. Key findings: (1) CHWs did not perceive AI as a threat (86%); (2) anticipated benefits—improved efficiency (92%), enhanced community respect (86%), upskilling (80%); (3) trust in AI (80%) alongside concerns about misdiagnosis (74%); (4) diverse data governance perspectives (58% government support); (5) limited privacy concerns (33%) despite vulnerabilities; (6) perceptual stability across scenarios demonstrated structural resilience of technological optimism in low-resource contexts. Conclusions: CHWs view AI-enabled health-environment integration as beneficial. This study contributes: (1) the Gender Paradox framework documenting differentiated adoption pathways with implications for gender-differentiated deployment; (2) structural resilience of technological optimism grounded in rational cost-benefit evaluation under scarcity. Deployment requires: algorithm transparency; gender-differentiated training; sustained technical support; and engagement with diverse values regarding privacy and community benefit.