Abstract:
Suppose that successively more powerful AIs are created, and they eventually perform most tasks cheaper and better than humans. This is likely to generate enormous wealth that can be used to fill human needs at one obvious cost: mass unemployment. In this paper, I will examine one possible response to this possibility, championed by theorists and practitioners alike. That is, to adapt our institutions to give people income and meaning in a world without work. I argue that given a plausible theory of political change I call realism, adaptation is a self-defeating strategy. According to realism, transformative change comes about if and only if the beneficiaries have the power to demand it. Replacing people with AI weakens their bargaining power and without power political elites will not adapt our institutions to give people income and meaning without work. The adaptationist project therefore fails, and we need to look to two other possible solutions, blocking and steering the development of AI.