Enhancing Software Engineering Education with Emotion-Aware AI Chatbots

Abstract

Despite the growing adoption of AI chatbots in engineering education, most research remains centered on English-speaking learners and overlooks the unique cultural, linguistic, and emotional needs of Arabic-speaking software engineering students. Additionally, many educational chatbots focus on delivering full solutions rather than guiding learners through critical thinking and problem-solving, which under mines engagement and cognitive growth. Emotional dynamics in engineering programming education, known for inducing stress and confusion, are also frequently neglected. This study ad dresses these gaps through a real-world controlled experiment involving 60 Arabic-speaking software engineering students assigned to one of four instructional setups: Text-Based inter action, Voice-Based interaction, Voice-Real Teacher interaction and Voice-Avatar interaction. Each participant completed the same programming task while emotional and learning-related outcomes were measured. Results showed that the avatar and voice-based setups yielded higher emotional stability, stronger engagement, and improved code quality. These findings demon strate that emotionally aware, culturally adaptive chatbot systems can significantly enhance engineering programming learning experiences, by supporting both cognitive and affective needs.