AI-Powered Assessment of Motor Development: Using Platforms Like KineticAI to Analyze Fundamental Movement Skills in Children

Abstract

The aim of this study is to examine the precision, dependability, and relevance of AI-based evaluations in contrast to conventional human evaluations. In all, 200 7–8-year-old students from urban and suburban schools participated in the study. Based on movement speed, accuracy, and smoothness, KineticAI’s assessment of their motor skills divided them into three categories: proficiency, developing, and emerging. A thorough examination of KineticAI’s validity and reliability was ensured by evaluating its psychometric qualities using COSMIN criteria. Furthermore, AI-generated scores and human evaluator ratings were compared using TGMD-3 as a standard. Mean Absolute Error (MAE), Intraclass Correlation Coefficients (ICC), and Bland-Altman plots were among the statistical techniques used to evaluate the degree of agreement. With an ICC of 0.94, the results show that KineticAI achieves great accuracy and dependability, showing strong consistency with human judgments. With running (3.8), jumping (4.2), hopping (5.1), and balancing (4.9) points, the AI system demonstrated a negligible mean absolute error (MAE) across motor skills, thereby proving its accuracy. Disparities in motor proficiency were also found by gender and school, with suburban girls scoring the lowest and urban boys the highest. These results highlight how crucial it is to provide everyone with fair access to organized physical activity programs to close developmental gaps. The study indicates that KineticAI offers a scalable, objective, and efficient alternative to traditional motor assessments. It is a valuable tool for use in schools, rehabilitation clinics, and sports training programs.