Rethinking AI strategy and policy as entangled super wicked problems

Abstract

This paper attempts a preliminary analysis of the general approach to AI strategy/policy research through the lens of wicked problems literature. Wicked problems are a class of social policy problems for which traditional methods of resolution fail. Super wicked problems refer to even more complex social policy problems, e.g. climate change. We first propose a hierarchy of three classes of AI strategy/policy problems, all wicked or super wicked problems. We next identify three independent super wicked problems in AI strategy/policy and propose that the most significant of these challenges - the development of safe and beneficial artificial general intelligence - to be significantly more complex and nuanced, thus posing a new degree of ‘wickedness.’ We then explore analysis and techniques for addressing wicked problems and super wicked problems. This leads to a discussion of the implications of these ideas on the problems of AI strategy/policy.