Yapay Zekâ Uygulamaları Perspektifinden DKAB Derslerinin Çoğulculuk Anlayışı
The Conception of Pluralism in RCE Courses from the Perspective of Artificial Intelligence Applications

Summary/Abstract: Purpose: Artificial Intelligence (AI) applications can reflect specific religious beliefs or interpretations depending on the datasets on which they are trained when generating knowledge about religious pluralism. As AI is increasingly incorporated into the domain of Religious Culture and Ethics (RCE) courses for purposes such as content creation, information retrieval, and the preparation of educational activities, its materials, its utilisation may potentially lead to issues such as the dissemination of misinformation, ethical violations, bias, and an inadequate representation of pluralistic perspectives. The objective of this study is to identify and evaluate the understanding of pluralism in RCE courses from the perspective of AI applications. It is significant in providing a technologybased perspective to religious pluralism debates, conducting comparative analyses, and contributing to identifying the potential benefits and limitations of AI in religious education. Method: The research was designed as a case study, and a structured interview form was employed to examine which religious education approaches and models are adopted in teaching Islam, other religions, and intra-Islamic interpretations. An experiment was conducted in which questions were posed to ChatGPT, DeepSeek, Gemini, Copilot, and Claude via two different IP addresses, on two separate occasions over a 20-day period. The responses were analyzed using content analysis and evaluated according to the following criteria: reliability, coherence, and consistency. Findings: It is evident that AI applications exhibited a paucity of depth in their approach to the subject matter. In contrast, DeepSeek and ChatGPT proffered a comparatively more expansive perspective. Although the outputs from DeepSeek, Gemini and Claude were generally accurate, some misinformation was present. With the exception of a single query in DeepSeek and Copilot, no sources were provided, and those cited were fabricated or inaccurate. The religious education approaches and models identified by most AI applications did not conceptually align with RCE curricula, whereas Gemini produced more compatible results. The religious education approaches and models identified in AI applications do not exhibit consistency with one another; rather, this variation appears to fluctuate depending on the IP address from which the query is made, as well as the temporal context and the specific content of the prompt. Conclusion: This study, which is distinguished by its original contribution in conducting a comparative analysis of the pluralism approach in RCE courses based on empirical data derived from five different AI applications, concludes that AI-generated data should be utilized only after being verified against core sources in the literature in terms of reliability, validity, and consistency. Furthermore, the findings suggest that responses to related queries may vary depending on differences in country, language, IP address, and temporal context; therefore, posing identical questions under varying conditions would enable comparative evaluation with the present study.

• Details
• Contents