Beyond digital access: co-designing inclusive artificial intelligence services with underserved Indonesian communities

Abstract

Background: Underserved communities in Kecamatan Paiton, Kabupaten Probolinggo, face unequal capacity to interpret and shape AI-assisted services despite living in a subdistrict of 67,709 residents within a regency where poverty reached 16.31% in 2025. Objective: This community engagement program aimed to co-design inclusive AI service scenarios with 30–50 community participants, including women microentrepreneurs, youth, village cadres, small farmers, pesantren/community learners, and coastal residents. Method: This program used a planning–implementation–monitoring-evaluation design supported by baseline surveys, attendance sheets, observation checklists, product rubrics, pre-post questionnaires, FGD guides, and partner validation forms. Results: Planning results were organized around 2–3 coordination meetings, 30–50 targeted needs-assessment responses, 4–6 stakeholder groups, and 3–5 prepared instruments. Implementation results were structured through 1 orientation session, 2–4 core workshops, 4–8 weeks of mentoring, ≥60–70% targeted output completion, and 5–8 documented AI service practice products. Implication: Monitoring-evaluation results used ≥75% attendance, ≥20% capacity improvement benchmarks, 1–2 validation forums, 1 institutional follow-up plan, and a 3–6 month sustainability review as evaluation indicators. Novelty: The novelty of this program lies in integrating participatory AI co-design, simple quantitative monitoring, community reflection, and institutional follow-up into a data-based community development model.