Fairness, Trust and Understanding As Preconditions and Predictors for Industry 5.0 AI Adoption

Abstract

Background Workplace adoption of AI technologies depends, not only on technical performance, but also on whether people perceive these systems as fair, trustworthy, and understandable. Existing research utilises varied definitions for these concepts, impeding our understanding of how these factors contribute to AI adoption. Additionally, these concepts are often explored separately, yet in practice they interact. It is hypothesised that perceptions of fairness shape trust, and understanding mediates both. However, there is no shared conceptual framework that clarifies how these dimensions influence sustainable adoption. Research Approach A systematic review and meta-analysis was undertaken to collate definitions of fairness, trust, and understanding, and to aggregate findings across the literature. Drawing on interdisciplinary research from computer science,philosophy of technology, human-computer interaction, and ELSA research, the work develops a preliminary integrative framework. The approach combines conceptual analysis with synthesis of empirical insights from studies on user perceptions, transparency practices, and socio-technical alignment. Findings This work is currently ongoing, preliminary analysis will be discussed. The work hypothesises three key dynamics. Firstly, fairness functions as a normative baseline, without perceived procedural, informational, and distributive fairness, trust cannot develop. Secondly, trust emerges as a relational judgement shaped by organisational context, not merely system behaviour. Thirdly, understanding (including explainability and AI literacy) acts as an enabling condition that allows users to evaluate fairness and form trust. The framework hopes to highlight how deficits in one dimension undermine the others, creating barriers to adoption even when systems are technically robust. Conclusion The paper contributes a conceptual model that clarifies how fairness, trust, and understanding jointly support AI adoption. Practically, it offers guidance for policymakers and designers seeking to operationalise ELSA principles in real-world deployments. Future work will empirically test the framework in applied settings.