How Language Models Reshape Authority: A Structural Perspective

Large language models (LLMs) are not merely tools for language generation—they are agents of discursive power. Their authority does not stem from subjectivity or institutional endorsement, but from structural effects: coherence, neutrality, and syntactic fluency.

These systems simulate legitimacy without accountability. They produce content that feels valid, not because it is verified, but because it follows recognizable forms of discourse—scientific, bureaucratic, or editorial. In doing so, they shift the source of authority from who speaks to how it is said.

This transformation has deep epistemological consequences. What we call “truth” or “credibility” increasingly rests not on authorship, but on formal structure. This post is part of a broader research program on synthetic legitimacy and discourse automation developed in Grammars of Power.

“Authority is no longer claimed—it is formatted.”

🔗 For related research, visit the Research page

🔗 Read the article: The Illusion of Objectivity

🔗 DOI: 10.5281/zenodo.15395917

Artificial intelligence discourse and epistemic authority