The Hidden Empire of AI Grammar: Power Without Politicians

How grammar, not algorithms, became the true operating system of power in predictive societies
About the Book

AI Syntactic Power and Legitimacy: How AI Structures Shape Power confronts a hidden dimension of artificial intelligence: power is no longer just exercised through policies, datasets, or algorithms, but through grammar itself.

The book shows how large language models establish authority not by what they mean, but by how they formulate. Passive voice, nominalizations, and non-referential clauses remove the human subject and create texts that look objective, neutral, and inevitable. In this regime, decisions appear as if they simply “exist,” while in reality they are executed by a system that no longer needs an author.

This work pushes beyond standard AI ethics discussions. Instead of asking whether AI is “biased” or “explainable,” it argues that AI produces legitimacy syntactically. Authority no longer requires a legislator. Once the grammatical form itself carries the command, sovereignty shifts from political will to structural execution.

*Why This Matters
*
Current debates on regulation and bias often miss the deeper transformation. If legitimacy itself is embedded in grammatical form, oversight mechanisms or ethics committees become secondary. By adopting AI-generated drafts, compliance reports, or automated policy texts, institutions already delegate their authority to an impersonal structure.

The consequences extend across domains:

• Law: Automated legal drafts obscure accountability by erasing the subject of responsibility.
• Finance: Compliance reports framed impersonally appear as obligations that “exist by themselves.”
• Healthcare: Clinical notes generated without agents reduce patients to passive objects.
• Governance: Policy texts compiled by AI present as objective, although their neutrality is only syntactic.

*Clear Examples for General Readers
*
To make this shift visible, here are cases where the form of the sentence changes authority itself:

1. Courtroom Example

Traditional: “The judge decided that the defendant is guilty.”

AI-shaped: “The defendant was found guilty.”
The subject disappears. Responsibility is obscured. The verdict looks like an inevitable outcome rather than a human judgment.

1. Hospital Record Example

Traditional: “Dr. Smith intubated the patient at 8:35 pm.”

AI-shaped: “The patient was intubated.”
Agency vanishes. If something goes wrong, the syntax leaves no actor accountable.

1. Corporate Compliance Example

Traditional: “The board requires compliance with safety protocols.”

AI-shaped: “Compliance with safety protocols is required.”
The obligation looks like a natural law rather than a decision. Authority dissolves into form.

1. Government Policy Example

Traditional: “The Ministry of Health mandates vaccination for children under five.”

AI-shaped: “Vaccination for children under five is mandated.”
The mandate becomes detached from the ministry. The sovereign body is erased.

1. Everyday Workplace Example

Traditional: “Your manager asked you to submit the report today.”

AI-shaped: “The report must be submitted today.”
No one asked. The command “exists.” You cannot question it. The authority is syntactic, not personal.

These examples show that the difference is not in meaning but in form. By erasing the subject, grammar constructs legitimacy as something neutral and non-negotiable. This is how AI silently naturalizes authority.

*Call to Action
*
The key question is no longer “What does AI say?” but “How does AI say it?” When syntax itself governs, authority becomes executable, impersonal, and detached from responsibility. Recognizing and resisting this hidden grammar of power is essential for technologists, policymakers, and citizens alike.

🔗 Explore the book here:

Official site: www.agustinvstartari.com/books

*Zenodo (full open version): *https://zenodo.org/records/17154108

*Academia.edu: *https://www.academia.edu/144039408/AI_Syntactic_Power_and_Legitimacy_How_AI_Structures_Shape_Power

*Author Information
*
Agustin V. Startari
Linguistic theorist and researcher in historical studies. Author of Grammars of Power, Executable Power, and The Grammar of Objectivity.

ORCID: 0000-0002-8727-4529

ResearcherID: K-5792-2016

Ethos

I do not use artificial intelligence to write what I don’t know. I use it to challenge what I do. I write to reclaim the voice in an age of automated neutrality. My work is not outsourced. It is authored.
— Agustin V. Startari