🧠 AI Ed Wiki

📄 Research Article

From Surface Learning to Deep Understanding: A Grounded AI Tutoring System for Moodle

◐ medium · 2026-05-07intelligent-tutoring, higher-ed, edtech-platform, llm, scaffolding, adaptive-learning-systems, metacognition, formative-assessment, adaptive-learning

Ostrowska, Kukla & Majstrak (2026) present an AI tutoring system integrated into the Moodle LMS designed to scaffold students from surface-level fact recall to deep conceptual understanding through adaptive questioning and feedback.

AI Tutoring for Moodle: Surface to Deep Learning

Core Contribution

Ostrowska, Kukla & Majstrak (2026) present an AI tutoring system integrated into the Moodle LMS designed to scaffold students from surface-level fact recall to deep conceptual understanding through adaptive questioning and feedback.

How It Works

The system operates within Moodle's existing infrastructure, using LLM-based tutoring to:

  • Detect the depth of student understanding from their responses
  • Adaptively scaffold toward deeper comprehension via Scaffolding techniques
  • Move students through Formative Assessment cycles that promote conceptual mastery

    • This grounded approach — embedding AI tutoring in an existing LMS rather than building standalone tools — addresses deployment barriers identified in the AI Tutor Effectiveness Review. Many intelligent tutoring systems fail to achieve real-world impact because they require new infrastructure; Moodle integration lowers the adoption threshold.

    Connections to the Wiki

    The system's focus on deep vs. surface learning connects to Metacognition research — students must recognize when they have only surface understanding. The adaptive approach aligns with Adaptive Learning but emphasizes qualitative shifts in understanding rather than quantitative difficulty adjustment. The Moodle deployment strategy echoes lessons from AI Peer Feedback Systems (AICoFe) about integrating AI tools into existing educational workflows.

    Unlike Tutoring Specific Vs General AI debates about specialized tutors, this system shows how general LLMs can be scaffolded into tutoring roles within familiar platforms. The focus on deep understanding complements Pedagogy AI Mistakes work on using AI errors for higher-order thinking.

    Open Questions

  • How does in-LMS tutoring compare to standalone systems in learning outcomes?
  • Can the deep learning scaffolding generalize across disciplines beyond the tested contexts?
  • What is the instructor's role when AI handles adaptive questioning within Moodle?

    • Connected Concepts

  • Scaffolding
  • Formative Assessment
  • Metacognition
  • Adaptive Learning

    • Connected Articles

  • AI Tutor Effectiveness Review
  • AI Peer Feedback Systems
  • Tutoring Specific Vs General AI
  • Pedagogy AI Mistakes

    • Citation

    Majstrak, A.A.O.M.K.G., Understanding:, F.S.L.T.D., Moodle, A.G.A.T.S.F., Skwarek, A.O.M.K.G.M.J.O.S.P.J., Wróblewska, A., Warsaw, F.O.M.A.I.S., scalability, R.G.L.E., & (LMS), M.S. (2026). From Surface Learning to Deep Understanding: A Grounded AI Tutoring System for Moodle