Can Artificial Intelligence Replace Traditional Public Health Disaster Management? A Critical Narrative Review

Abstract

Artificial intelligence (AI) has moved rapidly from experimental use to operational support in public health emergencies. It is now used, or actively proposed, for outbreak detection, forecasting, emergency triage, damage assessment, supply-chain planning, misinformation monitoring, and the management of emergency operations centres. This expansion has encouraged a provocative claim: that AI may eventually replace traditional public health disaster management. This critical narrative review examines that claim by comparing the strengths of AI with the complete set of responsibilities carried by public health and disaster-management institutions. The literature shows that AI can outperform conventional practice in selected, data-intensive tasks. It can process large and diverse information streams, recognise patterns that are difficult to detect manually, shorten reporting cycles, and support faster allocation of staff and supplies. Its value is greatest when information volume exceeds human processing capacity and when decisions must be repeatedly updated. However, technical performance in a narrow task is not equivalent to institutional replacement. Disasters create incomplete and unstable data, disrupt digital infrastructure, alter human behaviour, and magnify pre-existing inequalities. Models may fail under distribution shift, reproduce historical bias, or provide highly confident answers to questions that are partly ethical and political. Public health disaster management also depends on legal authority, field verification, professional judgment, community knowledge, trust, negotiation, and accountability. These functions cannot be transferred meaningfully to an algorithm. Evidence from COVID-19, natural hazards, humanitarian crises, and emergency care suggests that the most defensible future is neither traditional management without AI nor autonomous AI-led response. It is a human-led, AI-augmented system in which algorithms perform clearly bounded analytical and administrative work, while accountable professionals retain command, interpret uncertainty, protect rights, and make high-consequence decisions. AI can replace individual procedures; it cannot replace the public health disaster-management institution as a whole.