Dies ist eine Übersichtsseite mit Metadaten zu dieser wissenschaftlichen Arbeit. Der vollständige Artikel ist beim Verlag verfügbar.
CREATING ELT MATERIALS WITH CHATGPT: SERBIAN TEACHERS’ ATTITUDES AND PRACTICES
0
Zitationen
2
Autoren
2026
Jahr
Abstract
The integration of generative artificial intelligence tools such as ChatGPT into English language teaching is rapidly transforming instructional practices, yet empirical research on teachers’ attitudes and usage patterns remains limited, particularly in underrepresented contexts such as Serbia. This mixed-methods study investigates Serbian EFL teachers’ perceptions and practices regarding the use of ChatGPT for instructional material creation. Drawing on survey data from 53 in-service teachers and follow-up interviews with 10 participants, the research explores usage frequency, perceived usefulness, ease of use, benefits, challenges, and openness to future integration. Findings indicate that while ChatGPT is primarily employed for generating grammar and vocabulary exercises, lesson ideas, and early-stage planning, its use in supporting the four language skills and higher-order thinking remains limited. Perceived usefulness was positively correlated with usage frequency, and greater familiarity with ChatGPT predicted higher levels of integration. Paradoxically, more frequent users reported greater difficulty in prompt formulation, likely due to attempts at more advanced applications. Teachers with higher digital literacy and those in secondary education perceived ChatGPT as particularly beneficial. Most participants emphasized the need for formal training, underscoring the importance of targeted professional development. Overall, ChatGPT was viewed as a supplementary tool capable of enhancing efficiency, personalization, and innovation provided that its use is guided by pedagogical oversight and institutional support.
Ähnliche Arbeiten
Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI
2019 · 8.700 Zit.
Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead
2019 · 8.605 Zit.
High-performance medicine: the convergence of human and artificial intelligence
2018 · 8.133 Zit.
BioBERT: a pre-trained biomedical language representation model for biomedical text mining
2019 · 6.873 Zit.
Proceedings of the 19th International Joint Conference on Artificial Intelligence
2005 · 5.781 Zit.