Identifying Barriers to Transparent AI Governance in Human-Centered Work Design

Abstract

This study aimed to explore the organizational, sociotechnical, and cultural barriers to implementing transparent AI governance within human-centered work environments. A qualitative research design was employed, utilizing semi-structured interviews with 27 participants from diverse public and private sector organizations based in Tehran. Participants were selected using purposive sampling to capture a range of roles including management, IT, policy, and human resources. Interviews were conducted until theoretical saturation was achieved. Each interview was transcribed and analyzed using inductive qualitative content analysis. NVivo software facilitated systematic coding, enabling the emergence of main themes and subthemes from the data. The analysis yielded three overarching themes: structural and organizational constraints, sociotechnical misalignments, and cultural and cognitive barriers. Participants identified key challenges such as the absence of AI policy frameworks, siloed decision-making, and lack of ethical oversight mechanisms. Sociotechnical barriers included exclusion of stakeholders in AI development, technological opacity, and poor integration with human workflows. Cultural issues like distrust in AI systems, low AI literacy, and fear of job displacement further hindered transparent governance. These findings align with prior studies emphasizing the regulatory, ethical, and human-centered dimensions of AI transparency. The results demonstrate that transparency is not solely a technical feature but a multidimensional construct requiring organizational capacity, ethical culture, and inclusive governance processes. Transparent AI governance in human-centered work design is undermined by intersecting structural, technical, and cultural barriers. Overcoming these challenges requires a shift from compliance-based models to participatory and ethics-driven approaches that prioritize stakeholder engagement, organizational learning, and adaptive oversight. Addressing these barriers is essential for building trust, accountability, and fairness in the workplace as AI technologies become increasingly embedded in decision-making systems.