Large Language Models for Population-Level Public Health Communication: A Scoping Review

Fuente: Zenodo
Salvato in:
Dettagli Bibliografici
Autore principale: Farquhar, Hayden
Natura: Recurso digital
Lingua:inglese
Pubblicazione: Zenodo 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866902170202275840
author Farquhar, Hayden
author_facet Farquhar, Hayden
contents <p>This scoping review maps 552 studies applying large language models (LLMs) to population-level public health communication, following JBI methodology and PRISMA-ScR guidelines. Six databases and grey literature sources were searched (2019-2026). AI-assisted screening was validated against human review (Cohen's kappa = 0.83). Misinformation detection dominated (n = 186, 33.7%), yet misinformation correction was nearly absent (n = 11)---a 17:1 ratio. Only 9.6% of tools were deployed in practice, 83% of studies targeted English only, and 76% discussed no ethical considerations. The field has invested heavily in identifying health misinformation but almost nothing in responding to it. Redirecting effort toward correction, deployment, multilingual development, and non-communicable disease applications is essential. </p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_19439136
institution Zenodo
language eng
publishDate 2026
publisher Zenodo
record_format zenodo
spellingShingle Large Language Models for Population-Level Public Health Communication: A Scoping Review
Farquhar, Hayden
health communication
health literacy
large language models
misinformation
natural language processing
public health
scoping review
<p>This scoping review maps 552 studies applying large language models (LLMs) to population-level public health communication, following JBI methodology and PRISMA-ScR guidelines. Six databases and grey literature sources were searched (2019-2026). AI-assisted screening was validated against human review (Cohen's kappa = 0.83). Misinformation detection dominated (n = 186, 33.7%), yet misinformation correction was nearly absent (n = 11)---a 17:1 ratio. Only 9.6% of tools were deployed in practice, 83% of studies targeted English only, and 76% discussed no ethical considerations. The field has invested heavily in identifying health misinformation but almost nothing in responding to it. Redirecting effort toward correction, deployment, multilingual development, and non-communicable disease applications is essential. </p>
title Large Language Models for Population-Level Public Health Communication: A Scoping Review
topic health communication
health literacy
large language models
misinformation
natural language processing
public health
scoping review
url https://doi.org/10.5281/zenodo.19439136