Guided Learning with an AI Tutor: Pedagogical Support in Introductory Programming

Abstract

This innovative practice WIP paper describes MENTOR, a Machine-Enhanced Tutor for ORientation in Programming designed to support students in introductory programming courses. The system interacts with students through natural language, using conceptual explanations, step-by-step hints, and feedback to guide users in solutions, while intentionally avoiding the direct generation of complete solutions. The system has been developed for use in an introductory programming course at the Federal University of Mato Grosso do Sul (UFMS), focusing on Python. MENTOR was evaluated with 130 undergraduate computing students through a structured protocol combining a profile questionnaire, a Python-based technical assessment, and a multidimensional user experience survey. Preliminary results indicate that MENTOR acts as a catalyst for active learning, enabling students to transition from passive code consumption to structured problem-solving thinking.