Employee Perceptions of Artificial Intelligence Fairness in Software Development Organizations: Toward an Integrative Conceptual Model of Algorithmic Trust and Organizational Justice

Abstract

Software organizations are among the most intensive adopters of workplace artificial intelligence (AI), embedding intelligent systems directly inside technical work through AI-assisted code generation, automated code review, agentic coding assistants, and algorithmic productivity dashboards. A substantial body of research has examined how employees perceive the fairness of AI-based decisions in administrative human-resource contexts such as hiring, performance appraisal, and career development, and this literature consistently shows that fairness perceptions, not technical accuracy alone, determine whether employees accept, resist, or feel alienated by algorithmic decisions. However, comparatively little conceptual work has addressed how software developers specifically perceive the fairness of AI systems embedded in their core technical craft. This paper synthesizes organizational justice theory, algorithmic fairness research, and the software-engineering trust literature to identify four interrelated gaps: the concentration of fairness research in administrative rather than technical knowledge-work contexts, the inconsistent conceptualization of trust in software engineering research, insufficient attention to contextual and experience-level variation among developers, and the absence of an integrative model linking organizational justice to algorithmic trust and professional outcomes in technical settings. Building on Colquitt's four-dimensional justice framework, this paper proposes an integrative model in which Perceived AI Fairness in Technical Work, comprising distributive, procedural, interpersonal, and informational dimensions, shapes technical and relational algorithmic trust, which in turn influences developer engagement, professional identity, knowledge-sharing behavior, and turnover intention. AI transparency, organizational communication climate, developer experience level, and ethical leadership are proposed as boundary conditions. Six propositions are advanced to guide future empirical validation, including quantitative survey designs suitable for PLS-SEM analysis. The paper contributes a theoretically grounded and testable framework for understanding fairness perceptions in AI-augmented technical work, and offers practical guidance for software organizations seeking to implement AI in ways developers experience as fair and trustworthy.