“Risk för kognitiv överbelastning” – En kritisk innehållsanalys av generativ AI som ett didaktiskt verktyg i religionskunskap

Sammanfattning

This thesis aims to research how generative AI (GenAI) can be used as a didactic tool in religious education. The objectives of the thesis are to answer the following questions: How are Judaism, Christianity, and Islam portrayed in texts and images through GenAI?, What consequences could these representations have for religious education about Judaism, Christianity, and Islam?, and What are the opportunities and challenges with GenAI as a didactic tool in religious education? Previous research shows that traditional teaching materials often portray religions in stereotypical, simplified ways. Similar findings can be shown in images and texts generated through GenAI. The material in this thesis is the result of prompting the GenAI tool Microsoft Copilot to generate study questions, their answers, and one picture per religion. These are analyzed with a qualitative semiotic image analysis and a mixed method critical content analysis through the theoretical frameworks critical perspective and Mediated Learning Experience (MLE). The semiotic image analysis shows that Judaism and Islam are portrayed with inaccuracies and prioritize heteronormativity, while Christianity prioritizes prevalence. The critical content analysis shows that GenAI is unable to generate complex and in-depth answers and therefore reproduces stereotypes. The overall impression of both texts and images is that Christianity is portrayed as more favorable, followed by Judaism, and lastly Islam. This means, through MLE, that GenAI can not be used as a didactic tool without educating teachers and students about AI-literacy.