Urban decision-making is not so much a matter of intelligence, artificial or otherwise. A discussion of AI starting from cost-benefit analysis

It has long been assumed that effective decision-making presupposes the presence of a rational decision-maker and reliable information about alternatives, preferences and uncertainty. Unfortunately, the reality is quite different, particularly when one considers public decision-making in urban areas. In other words, and according to Herbert Simon's well-known formula, no decision-maker can operate without a form of “bounded rationality”. This is one of the reasons why traditional Cost-Benefit Analysis received so many criticisms, and other decision-support techniques were developed with the intention of taking our limited rationality more directly into account (e.g. Multicriteria Decision Analysis and Problem Structuring Methods). However, now that rationality no longer seems inherently constrained thanks to the advent of Artificial Intelligence, it seems reasonable to assume that decision-making processes themselves may undergo significant changes. Specifically, the research question that this article addresses is this: “What contribution can AI make to overcoming the limitations of certain methodologies used to support public decision-making in urban contexts?”. This question will be addressed by critically examining one of the main techniques of evaluation, Cost-Benefit Analysis, with particular regard to the pragmatic dimension of choice. • Certain evaluation techniques assume “bounded rationality” • AI seems able to overcome bounded rationality • The article critically discusses this possibility with a focus on CBA