Building health-literate artificial intelligence

This Perspective asks whether systems mediated by artificial intelligence (AI) can support health communication and health literacy rather than shifting interpretive burden onto users. We define health-literate AI as AI designed to align information, guidance and responsibility with users' abilities, contexts and needs. We propose four interrelated components: comprehension, agency, accountability and proportionality. These components help distinguish systems that are merely accurate or explainable from systems designed to support users in understanding what matters, what uncertainties remain and what to do next. We argue that the principles of health-literate AI should inform design, evaluation, research and governance across health-relevant AI ecosystems, so that innovation supports understanding, informed action and institutional responsibility.