Title: Impact of AI tutors on human learning behavior: A cross disciplinary analysis
Abstract:
Background: Learning experiences, nowadays, are being individualized and flexible by incorporating Artificial Intelligence (AI) tutors into classroom. Despite of its usage and implementation in various fields, a comprehensive multidisciplinary research is required to understand the impact of AI- based app on the human learning behaviour including motivation, cognition and emotion. The aim of this research is to assess the effect of AI-based tutoring systems on human learning behaviour focusing on individuals’ psychology, cognitive behaviour and motivation.
Methods: Data were gathered using a mixed-method approach from 120 students enrolled in CBSE-Board-affiliated schools. The NASA-TLX and Academic Motivation Scale (AMS Scale) questionnaires were used to gather baseline data. For 12 weeks, they were instructed to use AI-tutor. Data about the student's motivation, engagement, and learning experience with the AI-Tutor was gathered utilizing the Weekly Learner feedback tool. Behavioral analytics, performance indicators, and standardized psychological exams were used to examine changes in engagement, motivation, cognitive load, and retention. SPSS (version 22.0) was used to further analyze the gathered data. Pre- and post-intervention scores were compared using paired t-tests, and the predictive power of AI engagement indicators on learning outcomes was investigated using regression analysis.
Results: Overall, cognitive workload was significantly reduced throughout the study period, with mean NASA-TLX scores falling from 72.0 in Week 1 to 48.0 in Week 12. Intrinsic motivation increased from a mean of 4.2 to 5.8 and the extrinsic motivation increased from a mean of 4.9 to 5.5. Overall, the amotivation score was reduced from 3.1 to 2.2 after intervention. Learning experience showed significant gains across four categories: perceived understanding (3.2 to 4.3), personalization (3.1 to 4.5), feedback utility (3.4 to 4.6), and suitable pace of learning (3.3 to 4.2). Motivation related measures increased, with curiosity moving from 3.1 to 4.4, task motivation rising from 3.3 to 4.2, accomplishment increasing from 3.0 to 4.3, and AI preference changed from 2.9 to 3.9. Engagement indicators also increased, with focus (3.2 – 4.3), session completion (3.4 – 4.4), active interaction (3.5 – 4.5) and recommendation scores (3.1 – 4.3). The scores for the learning experience domain increased from 3.3 to 4.4, for motivation 3.2 to 4.3, and for engagement 3.3 to 4.4.
Conclusion: AI tutoring systems significantly enhanced learner motivation, engagement, and overall learning experience while reducing cognitive workload. The findings support the effectiveness of AI-enabled personalized learning environments; however, further development should incorporate emotional responsiveness and human-centered design principles to optimize long-term educational outcomes.
