Economic geographies of AI: Predicting, prescribing and programming economic futures

We are currently confronting the latest digital technology that is potentially reshaping Economic Geography—the so-called AI revolution. To date, there is relatively scattered economic-geographical research on AI as an industry in itself and even less research on AI’s role as a constituent of economic processes. Our goal with this paper is to (1) outline key aspects of AI’s possible impact on economic processes and (2) propose a research agenda that introduces an economic-geographical “way of seeing” to research on AI in the social sciences. Specifically, we propose a tripartite framework that considers AI’s role in predicting (non-neutral claims AI makes about the future), prescribing (delimiting acceptable actions), and programming (making concrete interventions with humans both “in-” and “out-” of “the-loop”) economic activities as particularly relevant to shaping future economic geographies. We apply this framework to three scales and domains of economic activity: everyday labour, geopolitics and global production networks, and subnational regional evolution. Centering the agentic power of AI within economic geography allows us to move from documenting AI’s spatial manifestations to better understand its effects on familiar concepts of agency, power, scale, networks and territorial configuration.