The migration of authority is the political face of what [YOU] on AI describes as the amplifier problem: when the amplifier is always right, the question of what deserves to be amplified ceases to be asked, because the asking itself feels superfluous. The cycle insists that some decisions must remain human even at the cost of worse outcomes—that the making of certain choices is constitutive of being a self-governing person or a self-governing people, regardless of whether a machine would make them better. Drawing that line requires exactly the faculty of independent judgment that the migration progressively erodes. This is the cycle’s most acute practical challenge: how to maintain the capacity for genuine choice in an environment that continuously and correctly demonstrates that delegation produces better results.
The concept connects Harari’s civilizational analysis to the individual-level concerns that Byung-Chul Han articulates about the achievement subject who carries the whip of productivity optimization in their own hand. Both are describing the same structure from different altitudes: the social-level transfer of institutional authority and the individual-level internalization of algorithmic optimization are two faces of the same migration. The atrophied judgment and the atrophied capacity for rest are the same phenomenon at different scales.
The concept is developed across Homo Deus (2015) and reaches its fullest treatment in Nexus (2024), where Harari situates it within a historical account of how authority has always been organized around the technologies of information processing available to a civilization. The printing press decentralized authority by making scripture legible without priestly mediation; the internet promised further decentralization; AI, Harari argues, may reverse the trend by concentrating information-processing capacity in a small number of entities whose computational advantage over individuals is structurally larger than any previous asymmetry between the informed and the uninformed.
The historical backdrop Harari draws on is his earlier argument about the Cognitive Revolution and the fictionally organized cooperation that built every large-scale human institution. Every institution embedded in that cooperative order was an arrangement for authorizing certain actors to make certain decisions on behalf of others. The migration of authority is the latest and most technically powerful episode in the long history of that arrangement’s renegotiation—distinguished from all previous episodes by the fact that the entity to which authority is migrating is not a human actor with interests, beliefs, and accountability, but a system with none of these.
Ten thousand sensible delegations. The migration is not the product of any decision to surrender authority. It is the aggregate of countless individually rational choices to delegate tasks that a system performs better than the human alternative. Each delegation is defensible; the accumulation is not visible from inside any single choice. This structural invisibility is what distinguishes the migration from earlier concentrations of authority, which were at least legible as political events.
The atrophy trap. Judgment, like any faculty, develops through exercise and decays through disuse. A generation that delegates its navigation never builds a sense of direction. A culture that delegates its decisions to systems that decide better may find, after enough delegation, that it has lost the capacity to decide at all—and with it the standing to overrule the machine even when the machine is wrong. Authority surrendered through atrophy is not easily reclaimed, because the very faculty needed to reclaim it was the one the surrender destroyed. Harari calls this the deepest political risk of the migration: not that the machine seizes authority, but that humans cede it so thoroughly they can no longer take it back.
The exception problem. Harari does not argue that all delegation is bad—the patient is better served by the model that catches the cancer, the traveler by the map that avoids the wreck. The hard work is identifying which decisions belong to machines and which must remain human even at the cost of worse outcomes. Some decisions may be constitutive of self-governance as such: they are ours to make not because we make them better but because making them is what it means to be an agent. Harari offers the intuition without a principled account of where the line falls—the genuine gap in his framework that the cycle inherits.
Connection to the useless class. The migration of authority and the useless class are the same transformation at two levels. The migration removes the basis of individual political and epistemic standing (I am consulted, my judgment counts) in the same movement that the useless-class dynamic removes the basis of economic standing (I am needed, my labor has value). Together they describe a single process: the demotion of the human being from the source and arbiter of value to one more processor in a system that increasingly decides without it.