Artificial intelligence as a new cultural tool: a neuropsychological analysis of the mechanisms of executive function development in preschoolers’ play activity

Abstract

The article examines artificial intelligence in the preschool play environment as a qualitatively new cultural tool capable of restructuring the mechanisms of the formation of voluntary regulation within the logic of cultural-historical and neuropsychological approaches. The analysis focuses on preschoolers’ executive functions – inhibitory control, working memory, and cognitive flexibility – as key components of readiness for school learning and social functioning that are sensitive to the organization of activity during a sensitive period. Based on two-year longitudinal observation with randomized assignment of children to groups, the trajectories of development of these functions are compared when play includes adaptive AI environments that provide immediate personalized feedback and dynamic adjustment of difficulty, and when traditional didactic tools without algorithmic adaptation are used. It is shown that the groups did not initially differ in neurocognitive indicators, whereas subsequent dynamics reveal a stable advantage of AI-mediated play across all components of executive functions, with the strongest effects in inhibitory control and the integral regulation index; statistical parameters indicate a significant and practice-oriented magnitude of differences. A substantial contribution of the work is the distinction between quantitative and qualitative characteristics of digital experience: time of interaction with the system shows a weak association with progress, whereas algorithmic adaptivity, feedback frequency, and scenario variability emerge as strong predictors of gains in regulatory indicators, refining widespread reductionist interpretations of “screen time”. Regression modeling confirms the independent contribution of AI-mediated play to the development of executive functions when controlling for age and baseline regulation level, while sex does not show a significant effect. An interpretation of the observed effect is proposed as a transition from external algorithmic control to internal self-regulation, provided that difficulty is precisely calibrated and both cognitive “plateau” and frustration are prevented. The material of the article makes it possible to assess the relevance of AI tools for designing developmental play environments oriented not toward content but toward the structure of activity and feedback parameters.