Integrating ChatGPT in intercultural education: Effects on students' performance in critical incidents exercises

Abstract

[EN] The rapid development of large language models like ChatGPT creates new possibilities for improving higher education. However, we still know little about how well it works for intercultural education. This study examines how generative AI, specifically ChatGPT, can be integrated into intercultural competence (IC) instruction through critical incident analysis. Conducted from March to June 2024 at a university in Shanghai, China, the intervention involved two groups of undergraduate English majors: a ChatGPT-assisted group (n = 25) and a non-ChatGPT group (n = 19). Both groups completed four textbook-based critical incident tasks during the semester, with ChatGPT support provided only to the ChatGPT-assisted group. Students’ written responses were qualitatively analyzed using open coding to identify patterns in terminology use, conflict identification, analytical depth, and problem-solving strategies. Findings indicate that the ChatGPT-assisted group demonstrated greater frequency and diversity of intercultural terminology and was more likely to propose preventive and long-term strategies. However, group differences in conflict identification and depth of analysis were limited, with greater variability observed in the ChatGPT group. The results suggest that ChatGPT can act as an effective cultural knowledge facilitator but may require complementary instructional design to foster deeper analytical skills and balanced IC development.