Pedagogy 2025

Abstract

Generative artificial intelligence (GenAI) is rapidly transforming assessment practices in educational and professional contexts, but little is currently known about how workers experience evaluation by AI and how evaluation influences a professional identity, ideologies of fairness, and forms of capital. In this article, I document a pilot AI-mediated assessment system designed for the recruitment division of a major IT company. The AI-mediated assessments combine hard-skill assessments (e.g., sourcing candidates) with AI-simulated soft-skill tasks (e.g., managing candidate objections), which are built from prompts written by senior recruiters in the recruitment division. Using semi-structured interview data with prompt creators (designers of the system) and junior recruiters (the anticipated users of the system), I rely on Darvin’s (2025) Identity-Ideology-Capital (IIC) framework to examine how individuals negotiate the legitimacy and implications of AI assessment. Results highlighted a design–use gap. Prompt creators framed AI literacy and prompt engineering as emerging symbolic and professional capital in their professional identity, positioning themselves as trainers of the machine and further investing in acquired technical knowledge. Junior recruiters, in contrast, positioned their professional identity in terms of embodied abilities such as persuasion, intuition, and charisma, concerned that these dimensions of their identity would not be recognized or valued by the AI. Members of both groups articulated a conditional trust in AI fairness and transparency, claiming to prefer a hybrid human-AI model to create a feeling of legitimacy. The study contributes to ongoing debates by illustrating how AI-mediated assessment continues to redistribute authority between humans and machines, thereby reconfiguring the symbolic economies of recognition and positions within an organization.