Evaluation of an AI-Based Feedback System for Enhancing Self-Regulated Learning in Digital Education Platforms

Abstract

The development of self-regulated learning (SRL) skills, including the ability to plan, monitor and reflect, is increasingly recognised as essential for academic success in online learning environments. Despite this, most digital learning platforms continue to provide limited feedback, typically focused on task outcomes rather than learning processes. This study investigates the effectiveness of an artificial intelligence-based feedback system integrated into a standard online course platform. The system delivers adaptive, process-oriented feedback by analysing anonymised engagement summaries and short reflective inputs, aiming to promote self-regulated learning strategies without requiring additional instructor involvement or manual input for feedback generation. A quasi-experimental study was conducted with 180 undergraduate students enrolled in a fully online course. Participants were pseudo-randomly assigned by an automated allocation script to an experimental group (n = 90) receiving AI-based adaptive feedback or a control group (n = 90) with standard LMS features. The system employed behavioural indicators (e.g., time-on-task, quiz activity and content engagement) and natural language analysis of reflective entries to generate personalised prompts related to goal-setting (including time management), effort regulation and metacognitive reflection. Data sources included post-course surveys, aggregated system interaction records, academic performance data and open-ended student feedback on the system’s perceived effectiveness and usability. Students who received adaptive feedback exhibited significantly stronger engagement with SRL behaviours, including earlier task initiation, increased use of optional learning resources and greater consistency in study routines. Qualitative responses indicated that participants found the feedback clear, timely, actionable and supportive of their cognitive and motivational processes. In contrast, control group participants primarily relied on grade-based feedback and exhibited fewer strategic adjustments during the course. The findings suggest that a lightweight, AI-driven feedback mechanism can be effectively integrated into online course platforms to support SRL at scale. This study demonstrates how adaptive AI feedback can meaningfully influence academic outcomes, planning behaviour and engagement with feedback in digital learning environments.

More by Galiya Zhusupkalieva

• Transforming engineering education in emerging economies: a comparative review of Kazakhstan and selected Sub-Saharan African contextsFrontiers in Education · 2026
• Distance learning using artificial intelligence: insights from educational analytics and student interaction dataFrontiers in Education · 2026
• The effectiveness of STEAM technologies on improving the professional competence of natural science studentsFrontiers in Education · 2025 · Citations: 14
• An in-depth framework for analyzing the outcomes of computer science instruction in an adaptive educational environmentInternational Journal of Innovative Research and Scientific Studies · 2025