Grounded in the TPACK framework, this study developed and validated an Artificial Intelligence Technological Pedagogical English Knowledge (AI TPEK) framework and measurement scale for university EFL teachers and examined the structural relationships among its knowledge domains. Based on valid questionnaire responses from 415 university EFL teachers, item analysis and confirmatory factor analysis (CFA)were conducted using SPSS 27.0 and AMOS 26.0 to evaluate the scales reliability and validity, while Smart PLS 4.1.1.2 was employed to examine the hierarchical relationships among the AITPEK domains. The findings indicate that:(1)the refined 28 item AI TPEK scale demonstrated good reliability, validity, and measurement stability;(2 )the traditional knowledge domains showed relatively weak correlations with the AI integrated knowledge domains, suggesting insufficient integration of technology, content, and pedagogy in EFL education; and (3)the structural model supports a transformative perspective of AI TPEK, indicating that the autonomous capabilities of AI are reshaping teachers’ knowledge development, whereby core knowledge is transformed into higher order integrative competence primarily through the mediating roles of cross domain integrative knowledge. These findings suggest that universities should promote the systematic integration of professional knowledge and AI technologies to enhance university EFL teachers’ digital literacy.