Evaluasi Desain Interaksi Aplikasi Pembelajaran Anak Berbasis AI Menggunakan Metode Child Computer Interaction

The development of Artificial Intelligence (AI) technology in education has encouraged the use of interactive learning applications, one of which is Duolingo, which utilizes AI to provide adaptive and personalized learning. This study aims to evaluate the interaction design of the Duolingo application using the Child-Computer Interaction (CCI) approach and to formulate improvement recommendations based on the evaluation results. The research involved 60 fifth-grade students of SD Negeri 129 Palembang as respondents through observation, usability testing, interviews, and a Smileyometer-scale questionnaire. Data were analyzed descriptively using average scores and percentages to assess the aspects of application appearance, usability, children's user experience, and feedback and interaction. The results showed a total score of 3,961 out of a maximum of 4,500, with an average of 4.40 and a percentage of 88.02%, categorized as very good. The highest percentage was obtained by the feedback and interaction indicator (88.78%), while the lowest was found in the ease-of-use indicator (87.50%). Students' open-ended responses indicated that the login process and menu navigation remained the main obstacles. Based on these findings, recommendations were formulated, including simplifying the login process, restructuring navigation, and developing features suited to children's needs so the application becomes easier to use, more engaging, and more effective in supporting learning.