What Develops When Artificial Intelligence Develops? Persona, Actor, or System
Abstract
Developmental accounts of intelligence ask how capacities emerge, stabilize, and reorganize over time. Artificial intelligence raises a prior question: what exactly is developing? This paper proposes the AGI Trilemma: persona, actor, and system as three competing ontologies for advanced AI. Persona describes the emergence of a coherent, socially legible identity; actor describes increasing goal-directed agency; system describes expanding computational organization without requiring either selfhood or centralized agency. These are not three stages on a single ladder. They are different accounts of what counts as intelligence and what development means. Contemporary systems may acquire cues from all three simultaneously. Developmental AI therefore risks importing human assumptions about self, agency, and competence into systems whose organization does not map cleanly onto any one of them. The trilemma offers a way to separate those trajectories before drawing conclusions about artificial development.