Learner Motivations in Human-AI Collaborative Reading: The AIR scale

This research examines learners’ motivations for using generative artificial intelligence (GenAI) during reading and introduces the GenAI-Assisted Reading (AIR) questionnaire. Although GenAI tools are increasingly integrated into everyday learning practices and can support efficiency and interaction with texts, concerns persist regarding inaccuracies, shallow processing, and overreliance. Existing accounts often portray students’ GenAI use as primarily effort-minimizing, potentially overlooking the diversity of motivations shaping human–AI collaboration in reading contexts. To address this gap, the present research developed and validated a multidimensional instrument for assessing learners’ motivations when engaging with GenAI while reading. Two studies were conducted. In Study 1 (N = 425), exploratory factor analysis identified four motivational dimensions: Task-Oriented motivation, reflecting strategic and cognitively engaged use of GenAI for reading tasks; Feel-Good motivation, capturing emotional and social drivers such as reduced frustration and increased enjoyment; Low-Effort motivation, reflecting the intention to minimize cognitive demands; and Translation motivation, representing the use of GenAI to support accessibility and cross-linguistic comprehension. Study 2 (N = 414) confirmed this structure through confirmatory factor analysis and demonstrated satisfactory reliability and construct validity. The AIR dimensions showed differentiated associations with related constructs. Task-Oriented and Feel-Good motivations were positively related to autonomy-oriented motivations for AI use, suggesting purposeful integration of GenAI into self-regulated learning. Low-Effort motivation was negatively associated with Need for Cognition, indicating greater reliance on GenAI for effort reduction among individuals less inclined toward cognitive challenge. Associations with technology acceptance variables further positioned GenAI-assisted reading within broader technology adoption processes, whereas minimal relations with task-switching tendencies and learning difficulties suggested that GenAI use is not merely compensatory. Overall, findings highlight the motivational heterogeneity underlying GenAI-assisted reading and provide a validated instrument for examining how motivational orientations shape constructive collaboration with AI.