Priesthood of Knowledge

In the AI Story

The AI industry of the twenty-first century exhibits priestly structure with a completeness exceeding any of Condorcet's examples. The concentration of knowledge is more extreme: the capacity to build frontier systems is possessed by a handful of organizations, located primarily in one country, funded by a small number of investors, controlled by a small number of individuals. The dependency is more comprehensive: users must trust that systems behave as advertised, that training data is representative, that failure modes are known — trust that is structurally identical to the medieval believer's relationship to the priest who read scripture on his behalf.

Condorcet's response was never reform of the priesthood but structural dissolution through universal education. He did not argue that priests should be made more virtuous. He argued that the laity should be made more knowledgeable. The goal was not better mediation but elimination of the need for mediation. This distinction is critical for AI governance. The dominant approach today is priestly reform — making builders more responsible, more transparent, more accountable. These measures are necessary but radically insufficient. Stewardship by the few, however well-regulated, remains dependency.

A feature of the contemporary situation Condorcet would have found particularly alarming: the opacity is not merely social but architectural. The medieval priest could, in principle, teach the believer to read scripture. The AI system's interpretive processes are not fully transparent even to the engineers who designed them. The black box problem means priestly mediation cannot be fully dissolved even if the will existed. This strengthens rather than weakens the case for universal evaluative literacy — the capacity to assess output reliability externally, without requiring full understanding of internal processes.

Regulation constrains priestly abuse. Education dissolves the structural dependency making abuse possible. Both are necessary. Education is more fundamental because it addresses not the behavior of the powerful but the capacity of the governed — the foundation on which, in every democratic theory from Condorcet to the present, legitimate governance rests.

Origin

Condorcet's anti-clericalism took definitive shape in his 1766 treatise on priests, and deepened through his collaboration with Voltaire on the Calas affair and the broader Enlightenment campaign against ecclesiastical authority.

The structural extension of the critique beyond literal clergy to all monopolies of specialized knowledge was Condorcet's distinctive contribution — and the element that makes the framework directly applicable to the AI industry.

Key Ideas

Structural, not individual. The attack is on the role, not the character of those filling it.

Dependency is the mechanism. Concentrated knowledge produces dependency; dependency produces power.

Dissolution, not reform. The goal is to make the priestly function unnecessary, not to appoint better priests.

Architectural opacity intensifies the problem. AI priesthoods may be more permanent than medieval ones because even priests cannot fully read their scripture.