中学教师AI认知

This study, based on a quantitative analysis of 221 Chinese middle school teachers, systematically explores the multi-dimensional mechanisms underlying the adoption of artificial intelligence (AI) educational technology. The key findings are as follows: (1) In terms of technology acceptance, perceived usefulness (β = .443) and ease of use (β = .353) collectively account for 59.3% of the variance in teaching cognition, thereby validating the applicability of the Technology Acceptance Model (TAM) in the context of educational AI; (2) Demographic factors exhibit selective influence, with ICT subject teachers demonstrating significantly higher teaching cognition than their non-ICT counterparts (ΔM = 0.82, p < .01), and male teachers showing stronger willingness to adopt AI (U = 3.21, p < .05). These results challenge the widely accepted "gender neutrality" conclusion in Western research; (3) The predictive mechanism of ethical cognition reveals unique characteristics, with 19.1% of its explanatory power primarily attributed to ease of use (β = .378) rather than risk perception (p > .05), indicating that middle school teachers prioritize the operational feasibility of AI over abstract ethical concerns. Theoretically, this study delineates the boundaries of TAM's applicability in an Eastern educational context while identifying cultural specificities in gender differences (U = 3.21) and subject background (ΔM = 0.82). Practically, it proposes a three-tiered development framework—demand-oriented curriculum design, low-threshold skill progression, and interdisciplinary collaboration communities—and innovates context-specific ethical training approaches to enhance the operationalization of ethical considerations and increase AI tool usage rates. Future research is encouraged to expand geographically into underdeveloped regions in central and western China and explore the integration of large-scale AI models into teaching assistants to further validate their impact on perceived usefulness. This study contributes new empirical evidence toward constructing a more universal theory of educational technology diffusion.

本研究基于对221名中国中学教师的量化分析，系统探究了人工智能（Artificial Intelligence，以下简称AI）教育技术采纳的多维度作用机制。核心研究结果如下：
(1) 在技术接受维度，感知有用性（β = .443）与感知易用性（β = .353）共同解释了教学认知59.3%的变异量，从而验证了技术接受模型（Technology Acceptance Model，以下简称TAM）在教育人工智能场景下的适用性；
(2) 人口统计学因素呈现选择性影响：信息与通信技术（Information and Communications Technology，以下简称ICT）学科教师的教学认知水平显著高于非ICT学科教师（ΔM = 0.82, p < .01），且男性教师对人工智能的采纳意愿更强（U = 3.21, p < .05）。上述结果对西方学界广为接受的“性别中立”结论构成了挑战；
(3) 伦理认知的预测机制呈现独特特征：其解释力的19.1%主要源自感知易用性（β = .378），而非风险感知（p > .05），这表明中学教师更重视人工智能的操作可行性，而非抽象的伦理顾虑。

理论层面，本研究厘清了技术接受模型在东方教育场景中的适用边界，并揭示了性别差异（U = 3.21）与学科背景（ΔM = 0.82）层面的文化特异性。实践层面，本研究提出了三级发展框架——需求导向的课程设计、低门槛技能进阶、跨学科协作共同体——并创新了适配特定场景的伦理培训路径，以强化伦理考量的可操作性，提升人工智能工具的使用率。

未来研究可拓展研究地域至中国中西部欠发达地区，并探索大规模人工智能模型与教学助手的融合路径，以进一步验证其对感知有用性的影响。本研究为构建更具普适性的教育技术扩散理论提供了新的实证依据。