Algorithmic Pedagogy in Early Childhood: Developing Adaptive Language Curricula for Whole-Child Social and Neuromotor Growth

Abstract

Digital learning tools in early childhood education often separate language learning from motor and social contexts. This study evaluates an Algorithmic Pedagogy Framework (APF) that uses adaptive artificial intelligence to connect real-time language tasks with physical motor feedback and peer social interaction. A 12-week randomized controlled trial involving 120 children in Sokoto state, aged 4–6 years compared traditional teacher-led physical play (Control A), static non-adaptive language applications (Control B), and the APF intervention (Experimental C). Outcomes included vocabulary retention, neuromotor precision, and peer turn-taking. The APF group demonstrated significantly greater active vocabulary retention (p < .001, ηp² = .22), higher gesture-task precision than the static-app group (d = .74), and greater spontaneous peer verbal exchange and physical turn-taking (d = .68). Adaptive language instruction can be designed as a facilitator of embodied, social learning rather than as a substitute for human interaction. The findings support a human-centred model in which AI operates mainly as an unobtrusive teacher co-pilot within active early-childhood environments.