["European Journal of Education, Volume 61, Issue 3, September 2026. ", "\nABSTRACT\nAn increasing bulk of research has recently focused on the role of artificial intelligence (AI) in second/foreign language (L2) education. However, few studies have examined users' perceptions about AI‐generated language and content in English as a foreign language (EFL) context. To fill this gap, the present study drew on the computers as social actors (CASA) framework and the technology acceptance model (TAM) to explore Chinese EFL teachers' perceptions about the authenticity and naturalness of AI‐generated language/content. The findings of thematic analysis indicated that the participants had perceived AI‐generated language as inauthentic and unnatural for eight reasons. Specifically, AI‐generated language was considered inauthentic because it is database‐oriented, ignores culture, emotions, semantics, and idiomatic expressions, does not produce genuine feelings in users, and lacks creativity and originality. Furthermore, the teachers saw AI‐generated language as unnatural because it lacks deep understanding and awareness, follows repetitive and predictable patterns, is too formal, and overlooks technicalities and idiosyncrasies. Each theme is discussed comprehensively, and implications are provided for teachers, students, teacher educators and AI designers to think deeper about the quality of texts and content produced by AI bots and chatbots before being injected into L2 classes.\n"]