AI Authoritarianism: A Government Without Constitutional Brakes - OpEd - Eurasia Review

TEXT START: The Trump administration’s cut of more than 270,000 federal workers alongside aggressive AI automation and the America.gov gateway is not smaller government but a shift to machines that execute decisions without human hesitation, objection or whistleblowing.

The Dissection

This is a constitutional alarm wrapped around a partisan account of administrative automation. Its strongest insight is that fewer government employees can produce greater state capacity when software scales surveillance, eligibility decisions, investigation, enforcement and targeting.

But the text sees the machine and misses the regime producing it. It treats AI government as an authoritarian temptation attributable to officials and technology companies, rather than as a structural consequence of competitive automation, fiscal pressure and concentrated control of AI infrastructure. It identifies the removal of human friction without analyzing who inherits the power released by that removal.

The Core Fallacy

The text confuses human presence with constitutional restraint. Bureaucrats are not constitutional brakes by default; they are often instruments of policy, and their employment does not guarantee conscience, resistance or accountability. Conversely, legal safeguards can theoretically be built into automated systems through auditability, appeal rights, liability and judicial oversight.

The deeper danger is not simply that machines do not hesitate. It is that authority, ownership and liability become separated. Private firms build the systems, officials deploy them, algorithms produce the decisions, and the public is left with no clearly accountable actor. Under P1 and P2, “human in the loop” is likely to become expensive ceremonial cover as competitive pressure pushes assistance toward recommendation and then delegation.

The article therefore correctly describes the trajectory but misidentifies its cause. AI authoritarianism is not merely a bad policy choice. It is the political form taken when machine capacity expands faster than institutional control and economic ownership remains concentrated.

Hidden Assumptions

• Human review is inherently more lawful, merciful or accountable than automated review.
• Constitutional institutions can preserve meaningful restraint after the state’s operational capacity becomes effectively limitless.
• The government can remain dependent on private AI firms without surrendering practical control to them.
• Future administrations will be able or willing to dismantle embedded databases, gateways and surveillance systems.
• Automation can be analyzed separately from labor displacement, ownership concentration and the collapse of productive participation.
• The primary threat is presidential overreach, rather than a permanent state-capital infrastructure inherited by every administration.
• Efficiency is an optional temptation, when competitive and fiscal mechanics make it increasingly compulsory.

Social Function

Partial truth, transition management and elite self-exoneration. The text accurately warns that automation multiplies state power and creates accountability gaps. It then retreats into familiar constitutional language, partisan blame and appeals to the Founders. That gives the audience a recognizable villain and a familiar remedy while avoiding the harder question: who owns the compute, data, models and distribution systems once human labor is no longer the main source of administrative capacity?

It is a warning label attached to the engine, not an analysis of the fuel. It does not confront the Discontinuity Thesis’s central break: when productive participation collapses, the population loses the economic leverage that once constrained both state and capital.

The Verdict

Accurate about the symptom. Incomplete about the cause.

The text correctly sees that a smaller workforce can coexist with a vastly more powerful government machine. It fails by treating that development as a corruption of an otherwise stable order. Under P1, P2 and P3, human-only administration becomes competitively obsolete, institutional resistance becomes harder to coordinate, and machine governance becomes the cheapest way to surveil, classify, allocate, punish and manage a displaced population.

The decisive question is not whether AI replaces government. It is who controls the replacement system—and whether ordinary people remain Sovereigns, indispensable Servitors, or merely processed subjects. Without ownership and enforceable control of productive AI infrastructure, “bureaucracy without bureaucrats” is not smaller government. It is the administrative form of systemic obsolescence: power after the public has lost the leverage to stop it.