Measures of learner-generative ai relationships
Sung-Hee Jin
Abstract
• Development and validation of novel Learner-Generative AI Relationship Scale • Three-factor structure: Affective Intimacy, Cognitive Competence, and Social Flow • Scale validity proven through ChatGPT-based argumentative writing task • Comprehensive measurement tool for improving AI-enhanced educational systems As Artificial Intelligence (AI) becomes increasingly integrated into educational environments, understanding the relationship between learners and AI systems is crucial for optimizing learning outcomes. This study introduces and validates the Learner-Generative AI Relationship Scale, a novel instrument designed to measure the multifaceted nature of learner-AI relationship in educational settings. The scale was developed through a rigorous process involving literature review, expert reviews, and cognitive pre-testing. An exploratory factor analysis with 95 undergraduate students confirmed a three-factor structure: Affective Intimacy, Cognitive Competence, and Social Flow, each comprising three sub-factors. The scale demonstrated good internal consistency and construct validity. To establish concurrent and predictive validity, 75 participants completed an argumentative essay writing task using ChatGPT. Concurrent validity was established through significant correlations with measures of attitude toward AI and AI self-efficacy. Predictive validity was confirmed through regression analyses, which showed that the learner-generative AI relationship significantly predicted learning engagement, perceived cognitive effects, and perceived motivational effects in a ChatGPT-assisted argumentative writing task. This study addresses a critical gap in the literature by providing a comprehensive tool for measuring learner-AI relationships beyond mere interactions and attitudes. The learner-generative AI relationship scale offers researchers and educators a valuable instrument for understanding and improving AI-driven educational systems, potentially informing the design of more effective AI-enhanced learning experiences.