Understanding AI guilt: the development, pilot-testing, and validation of an instrument for students
Cecilia Ka Yuk Chan
Abstract
Abstract This study explores the concept of AI guilt, a psychological phenomenon where individuals feel guilt or moral discomfort when using generative AI tools, fearing negative perceptions from others or feeling disingenuous (Chan, 2024). The phenomenon has become increasingly relevant as AI tools gain prominence in educational contexts. This paper introduces the development, pilot-testing, and validation of an instrument designed to measure AI guilt among students. Data were collected from 121 secondary school participants at an AI teaching and learning expo. The instrument identifies three dimensions of AI guilt: perceived laziness or inauthenticity, fear of judgment, and identity and self-efficacy concerns. Principal Component Analysis (PCA) and Cronbach’s alpha were employed to refine the instrument, ensuring its reliability and validity. By understanding AI guilt, educators and policymakers can mitigate its psychological effects and promote ethical AI usage in education.