教师生成式人工智能依赖量表:开发与应用OA

Generative artificial intelligence(GenAI)can empower teaching and learning,yet it may also lead teachers to devel-op excessive dependency and bring potential risks.Scientifically assessing teachers'dependency on GenAI is important for identifying risks and guiding appropriate use.Guided by the Person-Affect-Cognition-Execution interaction model,this study constructed an ini-tial scale through literature review,semi-structured interviews,and qualitative coding,in combination with two rounds of Delphi ex-pert consultation.The scale was then revised and validated using item analysis,exploratory factor analysis,confirmatory factor analy-sis,measurement invariance testing,and reliability testing.Results showed that the final scale comprises five dimensions—cognitive offloading,behavioral habituation,emotional attachment,environmental conformity,and adverse-outcome derivation—with a total of 23 items.The scale demonstrated good structural validity,convergent validity,and discriminant validity,and the reliability indicator met acceptable standards.Moreover,the scale exhibited satisfactory measurement invariance across groups defined by gender,educa-tional level,and region.A descriptive survey based on the scale indicated that teachers'overall GenAI dependency was at a moder-ately low level,but individual differences were pronounced.In addition,significant differences in dependency levels were found across gender,educational level,and regional background.This study provides a valid instrument for evaluating teachers'GenAI dependency and offers evidence to inform differentiated,tiered intervention and strategies for promoting reasonable use.