Local transplantation, adaptation, and creation of AI models for public health policy

Abstract

This paper presents the Transplantation, Adaptation and Creation (TAC) framework, a method for assessing the localization of different elements of an AI system. This framework is applied in the public health context, notably to different types of models that were used during the COVID-19 pandemic. The framework aims to guide AI for public health developers and public health officials in conceptualizing model localization. The paper provides guidance justifying the importance of model localization, within a broader context of policy models, geopolitics and decolonization. It also suggests procedures for moving between the different elements in the framework, for example going from transplantation to adapation, and from adaptation to creation. This paper is submitted as part of a special research topic entitled: A digitally-enabled, science-based global pandemic preparedness and response scheme: how ready are we for the next pandemic?