The Tragedy of Voluntary Servitude: a game-theoretic model of cognitive dependency in the age of artificial intelligence

The Tragedy of the Commons is among the most studied strategic dilemmas in the social sciences. I propose a new variant—the Tragedy of Voluntary Servitude—to model cognitive dependency on large language models and AI systems. Unlike the classical commons, where a shared natural resource is depleted by individually rational overuse, the resource depleted here is distributed cognitive and economic power across a population. Players choose to either Submit and accept deep AI dependency or Resist and maintain independent cognitive capability. I demonstrate that Submit is a dominant strategy for every player, and that this holds regardless of how individuals weight autonomy against productivity. Universal Submission emerges as the unique Nash equilibrium even when every player would prefer Universal Resistance. The result is power concentration in the AI provider, with gains in material productivity rendering the servitude comfortable and self-sustaining. Whether this equilibrium constitutes a tragedy depends on the welfare framework applied; I defend the position that cognitive autonomy is constitutive of meaningful agency, not freely substitutable for material gains. This game differs from classical commons in several respects: the damage is invisible to individuals, lock-in is irreversible, there is no moment of crisis, and subjects actively defend their own servitude. I review the historical toolkit for escaping analogous collective-action traps and examine why those tools may be insufficient here. I conclude by calling for economists, philosophers, and technologists to take seriously the prospect that we may be constructing a cognitive dependency from which there is no return.