Why We Created AI & Power Discourse Quarterly

AI & Power Discourse Quarterly was created from a simple observation: the public debate about artificial intelligence has expanded faster than the analytical vocabulary available to describe its political effects. Much of the discussion remains organized around performance, safety, bias, productivity, and regulation. These questions are necessary, but they do not fully explain what happens when automated systems begin to participate in the production of institutional language itself.

Language is not an incidental layer added after a technical decision. In law, administration, media, finance, security, and platform governance, language is often the form through which decisions become intelligible, legitimate, and executable. When AI systems generate, summarize, classify, rank, moderate, or reformulate that language, they enter directly into the architecture of authority.

Why discourse matters

A sentence can distribute agency. A passive construction can obscure responsibility. A nominalization can transform an action into an event without a visible actor. A risk category can convert a political subject into an administrative object. A recommendation can narrow available choices without appearing as a command. These are not merely stylistic effects. Repeated at institutional scale, they can alter how power is perceived and exercised.

This is the space the journal is designed to examine. Its object is not AI in isolation, and not discourse in isolation, but the point at which automated language becomes structurally consequential: where syntax, classification, interface design, moderation, and institutional procedure begin to shape the visibility of actors and the distribution of responsibility.

Beyond bias as the only framework

Bias remains a central concern in AI research, but it is not a sufficient theory of power. A system may avoid overtly discriminatory language and still reproduce asymmetry through form. It may treat one actor as capable of decision and another primarily as an object of humanitarian concern, security management, risk assessment, or intervention. It may present institutional preferences as neutral defaults. It may remove explicit coercion from the sentence while preserving coercive effects in the procedure.

For that reason, the journal approaches AI discourse as something that can be analyzed formally. Agency, modality, passivization, nominalization, deontic structure, risk vocabulary, default scopes, and patterns of attribution can all be studied systematically. The goal is to move from broad suspicion toward reproducible analysis.

The editorial project

AI & Power Discourse Quarterly exists to provide a space for work that crosses disciplinary boundaries without abandoning methodological discipline. Linguistics, political theory, sociology, law, media studies, philosophy, history, and AI governance all contribute to the same emerging question: how does power change when discourse becomes partially automated?

The journal's purpose is not to treat every machine-generated sentence as political, nor to assume that automation produces a single ideological effect. Its purpose is to identify where formal patterns become recurrent enough to matter institutionally, and to make those patterns available for critical examination.