Not Only AI: Rethinking the Epistemic Role of Computer Technology in Modern Economics
AbstractThe contemporary hype around artificial intelligence has not escaped economics. We witness increasing interest in the advancements in machine learning among economists who believe that AI tools will improve empirical analysis and evidential basis of modern economics, a proposal that follows earlier similar attempts of behavioural and experimental economists (Mullainathan & Spiess 2017). However, there is an important fact missing in these debates and proposals for a reform of economic research: since its modern development in the mid-20th century, economics has largely been influenced by computer technology.
We learn from historians of economic thought and historians of science that the neoclassical economic approaches were advanced during the Cold War era in the US by scholars who worked in research environments supported by military agencies (Erickson et al. 2013). Digital computers were built in these research environments and new sciences, inspired by the emergence of information technologies, were thriving there: cybernetics, cognitive psychology, systems engineering, AI research (Pickering 1995). Economics was not immune to the transformation by the computer that most modern science underwent. The philosophical significance of these developments, however, remains understudied. For instance, there is only a limited attention paid in philosophy of economics to computer simulations in economic research (Reiss 2011, Lehtinen & Kuorikoski 2021).
The aim of my talk is to argue for the importance of a much deeper and extensive philosophical analysis of the epistemic importance of computer technology in producing research in modern economics. Since the topic is vast, I narrow down my focus to the analysis of the epistemological consequences of the links between the computer and important theories of modern economics: expected utility theory and prospect theory. I analyse in which way idealizations of these theories reflect computational constraints of computer technology and I elaborate on the philosophical implications of this analysis.