Abstract
Improving student engagement and comprehension in STEM education remains a persistent challenge, particularly when traditional problem-solving exercises fail to resonate with learners’ interests. This study explores how integrating Human-Computer Interaction (HCI) principles with AI-driven problem generation can enhance classroom learning outcomes by personalizing educational content. We address the problem of low student motivation by leveraging Large Language Models (LLMs) to create academic problems tailored to individual learners’ preferences, hobbies, and motivational drivers. Our approach positions the AI as an interactive agent that dynamically adapts problem contexts using HCI frameworks, fostering a more engaging learning experience. In a controlled classroom experiment, one group received standard textbook problems while another engaged with AI-generated, personalized scenarios-such as engineering challenges embedded in science fiction narratives. Both groups were assessed through pre- and post-activity surveys, classroom participation, and targeted evaluations. Preliminary findings indicate that personalized problem generation significantly boosts student motivation, conceptual understanding, and enthusiasm for learning. Participants also expressed interest in applying this method across other academic subjects. These results suggest that combining generative AI with HCI can create more inclusive, adaptive, and effective learning environments, laying the groundwork for future personalized education systems.