Validation of the Model of Ethical Requirements of Artificial Intelligence in Education

Abstract

The present study was conducted with the aim of validating a model of the ethical requirements of artificial intelligence (AI) in education. In terms of purpose, the research was applied; in terms of data type, it employed an exploratory mixed-methods design; and in terms of implementation, it was carried out using a paradigmatic grounded theory approach in the qualitative phase and a cross-sectional survey method in the quantitative phase. The qualitative participants consisted of experts and specialists in the fields of artificial intelligence and information technology in education, including university faculty members and thematic administrators within the education system. The sample was selected purposively using the snowball sampling method, and interviews were conducted with 23 participants until theoretical saturation was achieved. The quantitative population included all district and secondary school administrators in the city of Tehran, totaling 1,880 individuals. Based on Cochran’s formula, a sample size of 319 was determined and selected through stratified random sampling. Qualitative data were collected through fieldwork using semi-structured interviews, while quantitative data were gathered through fieldwork using a researcher-developed questionnaire. To establish the trustworthiness of the qualitative findings, triangulation methods—including dependability, credibility, confirmability, and transferability—were employed, indicating that the data possessed adequate validity. The validity of the quantitative instrument was assessed through face validity followed by content validity, with the content validity ratio (CVR) for each component exceeding 0.42. The reliability of the instrument was calculated using Cronbach’s alpha, with coefficients for all components exceeding 0.70. Qualitative data were analyzed using theoretical coding, and quantitative data were analyzed using confirmatory factor analysis. The results indicated that the model of ethical requirements of artificial intelligence in education comprised 22 axial codes and 100 open codes within six dimensions of the paradigmatic model, all of which were confirmed by experts in the qualitative phase. According to the findings, the core phenomenon of ethical requirements of AI in education included three axial codes and 13 open codes; the causal conditions comprised three axial codes and 15 open codes; the contextual conditions included three axial codes and 16 open codes; the intervening conditions consisted of three axial codes and 14 open codes; the strategies encompassed five axial codes and 23 open codes; and the consequences included five axial codes and 19 open codes. Finally, the developed model was evaluated in the quantitative phase through confirmatory factor analysis, and the model fit indices indicated confirmation of the model. By accurately identifying conditions, strategies, and consequences, the proposed model provides a realistic yet normative representation of AI ethics in education.