Modeling K-12 teachers’ continuance intention to use generative AI: influencing factors and a hybrid SEM-ANN approach

K-12 teachers are the primary implementers of intelligent technology education. Prior research has emphasized teachers' initial intentions to adopt generative AI, but successful implementation depends on sustained long-term use. Exploring teachers' continuance intention to use generative AI in educational work and its influencing factors is vital for the depth and sustainability of technology adoption. This study integrates the Expectation Confirmation Model of Information System Continuance with Technological Pedagogical Content Knowledge, adds trust, system quality, and content quality as variables, and develops a model of continuance intention to use generative AI. Using 1,220 valid questionnaires from China K-12 teachers, we conducted an empirical analysis that combined structural equation modeling and artificial neural networks. This mixed approach both tests linear hypotheses and uncovers unknown nonlinear relationships in the data, thereby better revealing complex influence mechanisms. The results indicate that (1) the proportion of teachers using generative AI is relatively high, and their continuance intention to use it is above the midline. (2) Teachers' continuance intention to use generative AI varies significantly by age group, with older teachers showing lower intention. (3) Perceived usefulness, satisfaction, and trust are the primary determinants, and perceived usefulness and satisfaction exert the larger total effects on continuance intention. (4) Teachers' personal and technological characteristics form important foundations for prediction, and teachers' Technological Pedagogical Content Knowledge exerts the largest total effect on continuance intention. This study both confirms the value of mixed methods in educational technology research and supplies precise empirical evidence to guide administrators in designing tiered training and technology adoption strategies for teachers across different age groups.