The Digital Leviathan: When the Algorithm Becomes Sovereign
Seyed Alireza Alhosseini We used to fear the surveillance state because a government could watch us. The next problem is more...
We used to fear the surveillance state because a government could watch us.
The next problem is more profound:
What happens when the system watching us can also classify us, predict us, and decide who should be watched next?
This is the question behind my new paper, “The Digital Leviathan: AI, Surveillance, and the Erosion of Political Freedom.”
The paper examines how AI transforms surveillance from a human-scale activity into an automated architecture of political power.
Traditional surveillance required analysts, investigators, engineers, and bureaucratic layers.
AI can compress those layers.
A single operator can potentially use an AI system to process enormous quantities of information, identify patterns, generate intelligence reports, classify individuals, and accelerate targeting decisions.
That is not merely faster surveillance.
It changes the architecture of political power.
From Panopticon to Algorithmicon
Bentham's Panopticon gave us a physical architecture of observation.
Foucault transformed it into a theory of disciplinary power.
The AI era introduces something different:
the Algorithmicon.
It is decentralized, invisible, continuously operating, and capable of processing information at a scale no human surveillance bureaucracy can match.
The tower disappears.
The decision boundary remains.
The Algorithm Is the Sovereign
This is where the argument becomes more radical.
Carl Schmitt defined sovereignty through the power to decide the exception.
But imagine a political system where an algorithmic classifier determines that someone's behavior is politically sensitive—and that classification becomes the trigger for state action.
Who made the sovereign decision?
The politician?
The intelligence officer?
The institution?
Or the model's decision boundary?
If politically consequential decisions migrate from identifiable human authorities into opaque computational systems, sovereignty has not simply been automated.
It has been relocated.
Sovereignty Miniaturization
One of the deeper concepts explored in the paper is sovereignty miniaturization.
Historically, large-scale political control required large institutions.
AI can compress institutional capacity.
A contractor with an AI assistant can perform work that previously required engineering teams. A small intelligence office can automate analysis that once required many analysts.
The result is a disturbing asymmetry:
the scale of political power increases while the number of humans directly exercising it decreases.
Power becomes smaller in personnel—but larger in reach.
When Refusal Disappears
Political freedom is not merely the absence of physical coercion.
It requires the meaningful possibility of refusal.
If citizens modify their behavior because an invisible scoring system might classify them as suspicious, repression can occur before any visible act of coercion.
Self-censorship becomes anticipatory compliance.
The deeper question is:
Can a person meaningfully consent to a political order if they cannot meaningfully imagine refusing it?
The Problem Is Bigger Than One AI Company
The cases examined in the paper reveal a structural problem.
Banning an abusive account may stop access to one commercial model.
It does not necessarily eliminate the software already created, the surveillance infrastructure, the underlying datasets, or the ability to migrate to another model.
The capability can move.
Therefore, AI surveillance cannot be governed exclusively at the model-provider layer.
It requires governance at multiple levels:
Model → State → International
Model developers need stronger misuse detection and threat intelligence.
States require independent oversight of government AI surveillance.
Internationally, AI-enabled surveillance capabilities may require new mechanisms for transparency, verification, and accountability.
The Real Question
The central question is no longer simply:
“How powerful will AI become?”
It is:
“What happens to political freedom when computational systems become part of the machinery that determines who is visible, suspicious, dangerous, or controllable?”
The Digital Leviathan is not simply a story about machines watching humans.
It is a story about the relocation of political judgment into computation.
And once sovereignty can be expressed as a decision boundary, we need a new vocabulary for power.
Perhaps the most important question for the AI age is therefore not:
Who rules?
But:
Who—or what—gets to decide?
Read the full paper
The Digital Leviathan: AI, Surveillance, and the Erosion of Political Freedom
Read the full paper on PhilPapers
Author: Seyed Alireza Alhosseini Almodarresieh