From Telehealth to Artificial Intelligence: Digital Health Technologies for Indigenous Communities—A Systematic Review of Diagnostic Access, Ethical Governance, and Sustainable Health Equity

Indigenous communities worldwide face persistent health inequities rooted in colonial histories, geographic remoteness, and structural exclusion from diagnostic services. Artificial intelligence (AI) and digital health technologies are promoted as instruments of equity; however, the conditions under which they support rather than reproduce inequities remain contested. Following PRISMA 2020, PRISMA-Equity, and SWiM, we searched 12 databases (PubMed, Scopus, Web of Science, Embase, IEEE Xplore, ACM, CINAHL, Cochrane, SciELO, LILACS, Dimensions, Google Scholar) in English, Portuguese, Spanish, and French, plus structured grey literature. From 969 screened records, 39 studies met the eligibility criteria and were stratified into three layers: global, Americas/Latin America, and Brazil/Northeast. Deep-learning tele-otology and diabetic retinopathy screening in Aboriginal Australian contexts, suicide-risk machine learning with Native American communities, edge-AI maternal care with Indigenous Guatemalan midwives, and federated stress classification under Te Mana Raraunga emerged as the most mature applications. Latin America, Brazil, and the Northeast semiarid region were almost entirely absent; Brazilian tele-ultrasound work in the São Francisco Valley with Truká and Fulni-ô peoples, alongside Amazonian initiatives on tele-ophthalmology, machine-learning prediction of tuberculosis and malaria, cervical cancer screening, and culturally adapted cognitive assessment, offers a regionally grounded counterpoint. Indigenous data sovereignty, cultural safety, and external validation remain underdeveloped. We propose a nine-domain Responsible AI and Digital Health Implementation Framework aligned with the 2030 Agenda.