Prompt-Mediated Expertise and Generative AI in Knowledge Work: Credibility, Inequality, and Institutional Legitimacy

Abstract

Generative AI has rapidly become part of everyday knowledge work across universities, media, administration, and other professional settings. This expansion matters because AI systems do not simply accelerate output production; they also reshape how competence, credibility, and institutional trust are socially recognized. The purpose of this article is to examine how prompt-mediated expertise reorganizes professional authority, hidden labor, and legitimacy in AI-assisted work environments. The article adopts a qualitative, theory-driven design grounded in the sociology of expertise, organizational analysis, and studies of socio-technical systems. It draws on academic literature, governance documents, institutional materials, and public discussions concerning generative AI, workplace adoption, authorship, and accountability. Analytical attention is directed to prompt literacy, verification labor, institutional approval, access inequality, and the redistribution of discretion across AI-assisted workflows. A comparative interpretive reading is used to clarify how credible outputs are socially stabilized rather than automatically produced by model fluency alone. Prompt-mediated expertise emerges as a distributed form of competence in which human review, institutional validation, and unequal access remain central to the authority of AI-generated outputs. The article concludes that generative AI transforms expertise less by replacing professionals than by reorganizing the social infrastructure through which knowledge becomes trustworthy. Its contribution lies in providing a sociological framework for understanding how prompts, verification work, and institutional legitimacy interact in the contemporary politics of expertise.