ABSTRACT Artificial Intelligence (AI), automation, and digitally mediated systems are transforming institutions across the public and private sectors, creating new opportunities for efficiency, scalability, and improved decision-making. However, the effectiveness of these technologies depends not only on their technical capabilities but also on the institutional environments in which they are deployed. This paper argues that the central challenge of AI-mediated institutions is not AI itself, but the design of the institutional systems that shape how AI is evaluated, governed, integrated, and used in practice. Institutions determine what information becomes visible, what behaviors are rewarded, what knowledge informs decision-making, and ultimately whether AI strengthens institutional effectiveness or amplifies existing organizational weaknesses. Adopting a human-centered systems perspective, the paper develops a conceptual framework that explains how institutional design influences organizational behavior and AI-mediated outcomes. It demonstrates that institutional evaluation mechanisms, incentive structures, hierarchical information filtering, and behavioral normalization loops collectively shape emergent institutional patterns that affect organizational coherence, contextual decision-making, institutional learning, and adaptive capacity. Across diverse organizational settings, these mechanisms may unintentionally reinforce fragmented communication, misaligned incentives, hidden workload, weakened human judgment, and overreliance on measurable performance indicators while overlooking critical operational context and institution-sustaining contributions. Building on this institutional analysis, the paper advances the Human-Centered Ethical Systems Framework (HCESF) as a human-centered institutional design framework for AI-mediated environments. Rather than viewing AI as a replacement for human expertise, the framework positions AI as a capability that augments human judgment, supports institutional learning, strengthens ethical governance, and enhances collaborative decision-making through intentionally designed institutional systems. The proposed design principles emphasize contextual evaluation, transparent information flows, adaptive governance, equitable recognition of human contributions, and meaningful integration of AI within organizational processes. The paper concludes that meaningful and sustainable Human–AI collaboration depends not only on intelligent technologies but also on institutions intentionally designed to support human judgment, organizational coherence, and continuous adaptation. By shifting attention from technological optimization to human-centered institutional design, the study contributes a theoretical foundation and practical framework for developing AI-mediated institutions that strengthen resilience, preserve human agency, and promote long-term institutional effectiveness.
Taha Khan (Wed,) studied this question.