Actionability and legitimacy of AI-generated public-health knowledge

Determine how AI-generated knowledge can become sufficiently reliable, legitimate, and actionable within routine public-health practice, particularly in low- and middle-income countries, given the epistemic, infrastructural, ethical, and institutional conditions affecting AI-enabled disease surveillance.

Background

The paper identifies an unresolved implementation problem concerning the conditions under which AI-enabled disease surveillance can function responsibly and effectively in resource-constrained public-health systems. Existing research commonly examines technical, infrastructural, ethical, and organizational barriers separately, leaving their collective influence on implementation readiness insufficiently integrated.

The authors connect the problem to four dimensions of implementation readiness: epistemic adequacy of surveillance data, equitable distribution of digital infrastructure, ethical governance of health data, and institutional legitimacy and public trust. The question is especially significant in low- and middle-income countries, where fragmented information systems, limited connectivity, weak governance mechanisms, and constrained institutional capacity may prevent AI-generated information from becoming usable public-health knowledge and action.