The Adoption of AI in Recruitment Processes: Ethical, Social, and Bias Implications in Tunisia’s SME and Tech Startup Sector

Purpose The rapid adoption of AI-powered recruitment in emerging economies often occurs within contexts of institutional weakness and economic precarity, challenging Western-centric assumptions about algorithmic fairness. This study investigates how perceptions of algorithmic justice and platform transparency influence candidate trust and well-being in Tunisia, and whether chronic economic insecurity attenuates these psychological pathways. Design/methodology/approach Drawing on Algorithmic Justice Theory (Colquitt, 2001) and the Job Demands–Resources (JD-R) model (Bakker & Demerouti, 2017), we propose a moderated mediation framework. We theorize economic precarity as a chronic demand that depletes the psychological resources necessary to benefit from fair algorithmic treatment. The model is tested using survey data from 420 Tunisian job candidates, with contextual insights from 85 HR professionals. Findings Algorithmic justice (β = .24, p  < .001) and transparency (β = .28, p  < .001) enhance well-being indirectly through trust. However, economic precarity significantly weakens the trust–well-being relationship (Δβ = –.21, p  < .001). Algorithmic opacity directly reduces well-being (β = –.26, p  < .001). Graduates from non-elite institutions face a 2.4× higher risk of algorithmic misrecognition, indicating a digital reproduction of social stratification. Practical implications For AI ethics to be meaningful in contexts like Tunisia, technical fairness must be coupled with structural economic security. We provide actionable recommendations for localized bias audits, enhanced transparency protocols, and context-sensitive policymaking. Originality/value This is the first empirical study to situate AI recruitment ethics within the institutional voids and economic precarity of a post-revolution North African economy. It demonstrates that justice perceptions are not universally beneficial but are contingent on macroeconomic stability, advancing a context-sensitive framework for ethical AI adoption in the Global South.