Rédiger un PREreview

The phenomenon of artificial intelligence (AI) use in educational settings has attracted in-creasing scholarly attention, although applicable empirical findings are sparse—and con-flicting. This study seeks to resolve the ambiguities surrounding AI in education through a methodological contribution, merging neutrosophic stance detection and fuzzy-set Quali-tative Comparative Analysis (fsQCA). Neutrosophic analysis allows for an explicit mod-eling of truth, uncertainty/indeterminacy, and falsity, while merging such findings through fsQCA creates a relative account of extant research findings. After assessing four causal hypotheses related to AI-based learning opportunities through the Consensus Me-ter, an investigatory survey with 25 university participants explored necessary conditions with respect to experiencing improvements in learning outcomes. The findings indicate that the digital divide is a necessary and sufficient condition for effective AI educational experiences. Additionally, necessity conditions emerge for AI feedback and usage of AI-based platforms; however, the effectiveness of those platforms generates high uncer-tainty. Ultimately, the neutrosophic-fsQCA framework provides a viable technique to synthesize ambiguous findings through a systematic approach. Empirically, results re-veal that all stakeholders involved in potential AI-based learning need to ensure digital equity and high-quality design for interactive experiences to enjoy successful integration of AI in education.