EXPERIENCES WITH THE ADOPTION OF ARTIFICIAL INTELLIGENCE IN PUBLIC INSTITUTIONS WITHIN THE ARAB SOCIETY IN ISRAEL: A QUALITATIVE STUDY OF ORGANIZATIONAL TRUST AND DECISION-MAKING DURING CRISES

Abstract

Artificial intelligence (AI) is reshaping public administration by enhancing efficiency, data processing, and decision-making capabilities. However, in minority governance contexts, AI implementation is not a purely technical process but a socio-organizational transformation shaped by institutional trust, linguistic accessibility, and contextual alignment. This study examines employees’ experiences with AI adoption in public institutions serving the Arab society in Israel, focusing on organizational trust and crisis-related decision-making. Using a qualitative research design, data were collected over six months through semi-structured interviews with senior municipal officials, focus groups with administrative and IT professionals, and institutional document analysis. The study includes 50 participants from Arab local authorities in Israel. Data were analyzed using inductive thematic analysis to identify patterns in organizational perceptions and practices related to AI integration. Findings indicate that AI systems improve administrative efficiency and partially mitigate chronic understaffing in resource-constrained institutions. However, adoption is constrained by structural barriers, including linguistic exclusion due to Hebrew- and English-dominant interfaces, algorithmic misalignment with local administrative contexts, and limited access to culturally adapted training. Organizational trust emerged as the central mediating factor shaping AI acceptance and use, strongly influenced by leadership transparency, communication practices, and perceived institutional fairness. In crisis situations, AI enhances the speed of information processing and coordination, yet exhibits “contextual blindness,” limiting its responsiveness to local socio-cultural dynamics and informal governance structures. This creates a structural tension between algorithmic efficiency and contextual intelligence, reinforcing the need for human-in-the-loop governance in public decision-making processes. The study contributes to digital governance and public administration literature by introducing two analytically grounded constructs—algorithmic misalignment and contextual blindness—derived from a structurally under-researched minority governance context. It further highlights organizational trust as a key explanatory mechanism in shaping AI adoption outcomes in unequal institutional environments. The findings offer implications for designing culturally responsive, linguistically inclusive, and trust-sensitive AI governance frameworks in public sector organizations.