AI drifting and converting emergency policies

The impact of Artificial Intelligence (AI) on emergency policies has been widely examined in digital government literature, focusing primarily on AI’s role in policy design and delivery and on its effects on policy outcomes. However, limited attention has been given to how AI directly impacts emergency policies’ change, that is, in which ways AI reshapes and steers the policies it mediates. Drawing on Hacker et al. (2015) framework of institutional change, particularly the concepts of drift and conversion, this research provides a novel lens to understand policy change in emergency contexts. Drift occurs when policies remain formally unchanged but fail to adapt to evolving contexts, altering their effects. Conversion refers to the redirection of policies towards new purposes without modifying their formal structure. By applying this framework, the paper investigates AI’s dual role in changing emergency policies, both as a source of resistance to formal policy change and a catalyst for policy redirection. The research adopts an explanatory case study of Peru’s welfare policies during the COVID-19 pandemic, focusing on the use of the SISFOH AI system to allocate emergency subsidies. Findings demonstrate how AI changes the policies it mediates by drifting formal policy modification and enabling conversion by reshaping policy ends. This paper contributes to the digital government literature by highlighting AI’s ambivalent role in emergency policy change and offering fresh insights into the intersection of AI and emergency policies.