Answers are grounded in this paper and cite their sources.
Amid the rapid global expansion of artificial intelligence (AI), AI education has become an increasingly integral component of K–12 educational systems. Teachers’ instructional intentions are instrumental in shaping classroom implementation and professional development trajectories; however, empirical research has paid limited attention to teachers’ perspectives within AI education contexts. In response to this research gap, the present study expands the Theory of Planned Behavior (TPB) by integrating AI literacy, facilitating conditions and AI anxiety, thereby developing a more integrative motivational framework for investigating teachers’ intentions to teach AI. Using cross-sectional survey data from 218 secondary school information technology teachers, covariance-based structural equation modeling was employed to test the hypothesized relationships. Machine learning classification analyses were additionally conducted to examine the predictive performance of the theoretically relevant variables. The extended TPB model explained 64.2% of the variance in teachers’ intentions to teach AI. Attitude, subjective norms, and perceived behavioral control were significantly and positively associated with behavioral intention. Subjective norms were positively associated with attitude, whereas AI anxiety was negatively associated with attitude. Although AI literacy was not directly associated with either attitude or perceived behavioral control, it was significantly associated with subjective norms and AI anxiety. Facilitating conditions were strongly positively associated with perceived behavioral control. The machine learning predictions indicated that statistically significant structural associations do not necessarily translate into stable individual-level prediction. The findings reveal a nuanced role of AI literacy: it was positively associated with subjective norms and AI anxiety but was not significantly associated with attitude or perceived behavioral control. The unexpected positive association between AI literacy and AI anxiety suggests a possible pattern of informed anxiety. The machine learning analyses further showed moderate to good discrimination in out-of-sample testing for intention, attitude, perceived behavioral control, and subjective norms, but near-chance performance for AI anxiety. This study highlights the need for teacher professional development initiatives that address not only AI-related knowledge and skills but also teachers’ emotional concerns and perceived environmental support.
No takes yet. Share an insight, caveat, or question.
Gu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: