Barriers to the Application of Artificial Intelligence for Performance Evaluation in SAMPAD Schools

Abstract

This study aimed to identify and conceptualize the barriers to applying artificial intelligence (AI) for performance evaluation in SAMPAD schools and to explain the human, organizational, technical, and ethical dimensions influencing the adoption and implementation of this technology. This qualitative study employed reflexive thematic analysis. Participants comprised 17 academic and executive experts with substantial experience in gifted education and the educational application of emerging technologies, recruited from Tehran, East Azerbaijan, West Azerbaijan, and Kurdistan provinces through maximum-variation purposive and snowball sampling. Sampling continued until thematic saturation was achieved. Data were collected through 45–75-minute semi-structured interviews and analyzed according to Braun and Clarke’s six-phase reflexive thematic analysis framework. Trustworthiness was enhanced through prolonged engagement with the data, participant triangulation, member checking, comprehensive documentation of the analytical process, reflexive memo writing, and independent coding. Analysis generated 214 initial codes, which were refined into 87 preliminary concepts and subsequently organized into four major themes and 16 subthemes. The four principal themes were attitudinal and cultural barriers, structural and organizational barriers, technical and infrastructural barriers, and ethical and legal barriers. Resistance to technology, concerns about displacement of the human role, distrust of AI-generated outputs, and traditional approaches to assessment interacted with bureaucracy, centralized decision-making, organizational inflexibility, and excessive workloads. Inadequate technological infrastructure, implementation and maintenance costs, bandwidth limitations, and insufficient technical support further constrained implementation. Ethical and legal concerns involved student privacy, data protection, algorithmic bias, lack of transparency, regulatory gaps, and unclear accountability. Divergent findings additionally highlighted reduced human interaction, digital inequality in disadvantaged regions, and limitations of Persian-language AI content. AI-based performance evaluation in SAMPAD schools is constrained by interconnected human, institutional, technical, and normative barriers. Effective implementation therefore requires an integrated strategy encompassing teacher capacity building and trust development, organizational reform, equitable infrastructure investment and sustained technical support, and transparent ethical and legal frameworks for data protection, bias mitigation, explainability, and accountability.