Artificial intelligence and digital health equity: a post-pandemic evidence synthesis and implementation safeguards framework

IntroductionThe rapid expansion of AI-enabled digital health after COVID-19 created new possibilities for extending care while raising concerns about unequal access, subgroup underperformance, and weak accountability. These effects remain difficult to interpret because telehealth infrastructure, predictive analytics, clinical decision support, and generative AI operate through different equity mechanisms.MethodsA transparent narrative evidence synthesis was conducted using peer-reviewed literature and selected governance documents. Structured searches of PubMed, Scopus, Web of Science, and IEEE Xplore were refreshed on March 5, 2026, for English-language sources published between January 1, 2020, and December 31, 2025. A second reviewer independently assessed a sample of 12 full-text inclusion and exclusion decisions using the same pre-specified criteria, with 100% agreement. The final corpus included 29 sources.ResultsEvidence was strongest for digital access barriers, subgroup underperformance, and implementation-related governance concerns. AI-assisted screening and digitally mediated remote support showed conditional benefits in targeted settings when high-touch support was built into implementation. Evidence on generative AI and language-support tools was comparatively smaller, newer, and concentrated in representational-bias applications, with limited evidence of sustained downstream equity gains.ConclusionAI-enabled digital health may reduce selected barriers to care when supported by equitable access conditions, subgroup validation, accountable governance, and implementation oversight. Equity gains were conditional rather than automatic, and evidentiary depth varied substantially across modalities. The review distills recurring patterns into a synthesis-derived operational roadmap for procurement, validation, implementation, and post-deployment monitoring.