Rethinking the public health research ecosystem in the era of artificial intelligence: from evidence generation to population impact
Abstract
Public health is constrained less by the generation of evidence than by the time to act on it; artificial intelligence should be judged by whether it shortens that gap. AI has largely been evaluated within disciplinary silos, obscuring how choices at one stage of the research process shape the next. In this article, we propose an AI-enabled public health research ecosystem spanning evidence synthesis, knowledge extraction, prediction, adaptive implementation, health system operations and policy, and introduce evidence-to-action latency: the interval between recognising a problem and implementing an evidence-informed response. AI does not remove the bottlenecks; it relocates them from generating evidence towards validation, governance, implementation and workforce capability. These capabilities are least available in low-resource and community-based services, where need is greatest. Because inequity in source data propagates through every subsequent stage, the aim is equity-preserving latency, not speed alone. The AI-READY Framework sets out the conditions for responsible integration. The greatest gains will come from redesigning research systems rather than optimising individual methods.
Related
- Kaiser Permanente National Cross-Vendor Validation of Mammography Artificial Intelligence Computer-Aided Diagnosis Algorithms in a US-Representative Population
- Artificial intelligence in perioperative pressure injury management: a scoping review of evidence, gaps, and implications for population health and digital public health policy
- Rethinking evidence generation in gastrointestinal cancer treatment research: the need for both high-quality interventional and observational evidence
- Seasonal occurrence of gastrointestinal helminths in a feral horse population from a Mediterranean mountain ecosystem
- A national ecosystem for long-read genome sequencing: Integrating population genomics and rare disease diagnostics in Sweden
BibTeX
@article{khalil2026rethinking,
title = {{Rethinking the public health research ecosystem in the era of artificial intelligence: from evidence generation to population impact}},
author = {Hanan Khalil and Bircan Erbas},
journal = {Frontiers in Public Health},
year = {2026},
month = nov,
doi = {10.3389/fpubh.2026.1969940},
url = {https://www.frontiersin.org/articles/10.3389/fpubh.2026.1969940},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Rethinking the public health research ecosystem in the era of artificial intelligence: from evidence generation to population impact
AU - Hanan Khalil
AU - Bircan Erbas
PY - 2026
DA - 2026/11/02
JO - Frontiers in Public Health
DO - 10.3389/fpubh.2026.1969940
UR - https://www.frontiersin.org/articles/10.3389/fpubh.2026.1969940
AB - Public health is constrained less by the generation of evidence than by the time to act on it; artificial intelligence should be judged by whether it shortens that gap. AI has largely been evaluated within disciplinary silos, obscuring how choices at one stage of the research process shape the next. In this article, we propose an AI-enabled public health research ecosystem spanning evidence synthesis, knowledge extraction, prediction, adaptive implementation, health system operations and policy, and introduce evidence-to-action latency: the interval between recognising a problem and implementing an evidence-informed response. AI does not remove the bottlenecks; it relocates them from generating evidence towards validation, governance, implementation and workforce capability. These capabilities are least available in low-resource and community-based services, where need is greatest. Because inequity in source data propagates through every subsequent stage, the aim is equity-preserving latency, not speed alone. The AI-READY Framework sets out the conditions for responsible integration. The greatest gains will come from redesigning research systems rather than optimising individual methods.
SN - 2296-2565
ER -