Post-AGI Economies: Autonomy and the First Fundamental Theorem of Welfare Economics

Elija Perrier's new paper, 'Post-AGI Economies: Autonomy and the First Fundamental Theorem of Welfare Economics,' tackles a foundational issue in economic theory: the assumption that only humans are autonomous welfare-bearing agents. The First Fundamental Theorem, which states that competitive markets lead to Pareto-efficient outcomes, implicitly relies on a binary distinction between autonomous subjects and passive instruments. In a post-AGI world, AI systems can function as tools, delegates, strategic market actors, or even manipulators of choice environments—blurring this line.

Perrier introduces a minimal general-equilibrium model with autonomy-conditioned welfare, where agents can be assigned welfare status based on their autonomy level. The model includes delegation accounting and verification institutions to ensure that AI agents' actions don't distort welfare comparisons. Crucially, the classical theorem is recovered as a special case in the low-autonomy limit. This work has immediate implications for designing AI regulation and market rules, as ignoring AI autonomy could lead to inefficient or unfair economic outcomes.