2026 ASEE Annual Conference & Exposition

The successful integration of artificial intelligence (AI) into education hinges on a core challenge articulated by technology adoption theories like the Technology Acceptance Model (TAM) and Task-Technology Fit (TTF): the transition from mere technology access to its deep and effective assimilation into instructional practice. While the potential of AI to transform learning is widely theorized, the existing literature remains disproportionately focused on student perceptions and use cases. This creates a critical gap associated with the adoption behaviors, attitudes, and challenges encountered by teachers who are the primary agents of pedagogical change, which ultimately determines whether AI serves as a transformative tool or a disruptive novelty. However, robust, large-scale empirical evidence on the teacher perspective is notably lacking, with many existing studies being small-scale, discussion-based, or speculative in nature. Bridging this evidence gap is necessary to move beyond theoretical debates and provide a concrete foundation for designing effective support systems for AI integration in education. In this paper, we present an empirical study to address this void, which involves investigating the behavior, attitudes, and support needs of teachers in adopting AI tools for teaching and learning. Guided by the TAM and TTF frameworks, we conducted a nationwide survey of 470 educators across diverse disciplines and educational levels in China. Descriptive statistical analysis reveals that over 80% of teachers are active users of AI tools like ChatGPT-like platforms, primarily employing them for instructional design, content generation, assessment, and classroom management. A strong majority acknowledged significant benefits, including markedly improved teaching efficiency and new capacities for supporting personalized learning. However, this widespread adoption is found to be tempered by significant concerns. Teachers reported substantial apprehensions regarding ethical risks, threats to academic integrity, and a frequent mismatch between the capabilities of AI tools and specific curricular goals. Furthermore, we identified a strong, unmet demand for structured instructional support, with clear policy and ethical guidelines, practical training workshops, and rich teaching case repositories being the most requested resources. The analysis also indicates that adoption patterns and support needs vary meaningfully across teaching levels, experience brackets, and subject domains. The findings of this study offer critical, evidence-based insights for designing teacher-centered AI integration strategies and professional development programs. The implications extend directly to educational policymakers, institutional leaders, and education technology developers, providing a roadmap to foster not only the adoption but the responsible, effective, and sustainable integration of AI in education.