The use of artificial intelligence (AI) technologies in second/foreign language education has recently gained a bulk of attention. However, the emotional experiences of English as a foreign language (EFL) learners in AI-mediated classes have been ignored. To fill this gap, the present qualitative study examined 34 Chinese EFL students' perceptions of AI-induced emotions and regulation strategies. A semi-structured interview and a narrative frame were used to collect the data. The gathered data were thematically analysed through the latest version of MAXQDA software (v. 2023). The findings revealed that Chinese EFL students had mostly experienced positive emotions of 'motivation', 'excitement', 'engagement' and 'confidence'. On the negative side, they reported experiencing 'frustration', 'anxiety' and 'stress' more frequently in their classes. Furthermore, the study indicated that the participants had used six strategies, namely 'seeking help from others', 'shifting attention', 'cognitive change', 'persistent practice', 'staying positive' and 'suppression' to regulate their AI-induced emotions. The findings are discussed and implications are provided for EFL students and educators to understand the emotional aspect of AI injection into L2 education.