<p>Teacher questioning serves as a crucial mechanism to foster learners’ higher-order thinking skills. In contemporary college education, teachers leverage interactive platforms to pose questions, and evaluate students’ comprehension of key concepts by collecting their responses, thereby managing the pace of instruction. However, the rise of AI chatbots raises concerns that students may over-rely on AI tools to generate answers, potentially impacting their mastery of knowledge and the accuracy of teacher assessments.We employ an interactive classroom at a Chinese university as a case study, and then explore whether the teacher questioning strategies will affect students’ tendency of using AI chatbots to generated answers, and whether students’ usage of AI chatbots will have impact on their final academic performance. First, we develop an AI plagiarism detection algorithm, leveraging prompt-learning and the pre-trained RoBERTa model, to accurately identify AI-generated answers. Then, we employ linear regression and chi-square test to investigate the correlation between teachers’ questioning strategies and students’ utilization of AI chatbots. Finaly, we calculate the Pearson correlation coefficient and the value of Cramer’s V to ascertain the association between the frequency of students’ AI chatbot usage and their final academic performance. Our proposed model can identify the AI-generated content accurately. Moreover, we observe a significant correlation between the types of questions posed by teachers and the frequency of students’ AI usage, particularly for questions that engage higher-order thinking skills. Notably, the correlation between AI usage frequency and students’ final grades is weak, suggesting the incorporation of AI tools in interactive learning environments may have a limited impact on their academic performance.</p>