When Résumés Become Instructions: Rhetorical Governance and Prompt Injection in AI-Mediated Hiring

As organizations increasingly use AI-assisted tools for hiring and candidate screening, résumés are now evaluated by both human reviewers and automated systems. This study examines how AI-assisted résumé screening responded to hidden instruction-like content (prompt injection) embedded within fictional résumés. Across three screening environments, prompt injection attempts produced limited influence on hiring outcomes, though detection became less consistent as governance structures were reduced. The findings suggest that AI-assisted résumé evaluation is shaped not only by computational systems, but also by rhetorical governance structures, including genre expectations, evaluative criteria, and professional communication conventions. Rhetorically strong résumés consistently outperformed manipulation attempts.