The Impact of Widespread Use of AI-Driven LLMS on Academic Self-Esteem, Loneliness and Cognitive Flexibility of Adolescents

Abstract

he idea of Artificial Intelligence (AI) comes from thinkers like Alan Turing, and the technology itself has permeated daily life in the form of large language models (LLMs). This paper examines the impact of this technology on adolescent development. Technology is widely available to this age group, unlike children; and unlike adults, their cognitive faculties are still developing. Hence, they are an important population to consider. The study examines attitudes toward technology - positive, negative and anxiety about dependence as mediators between AI Use and three outcome variables - academic-self-esteem, cognitive flexibility and Loneliness. The study uses a cross-sectional survey design collecting data from Indian school children (aged 12 to 19, N = 407). The study was pre-registered on OSF. Preliminary reliability analysis for the tools and correlations were conducted. Based on these findings, two mediation models were analysed. Positive attitude towards technology/AI was a full mediator between AI Use and Cognitive flexibility, and academic self-esteem was a partial mediator between cognitive flexibility and loneliness. We also found that anxiety about dependence had a small but significant predictive impact on loneliness through regression analysis. The study has important implications for the study of the impact of AI, as it examines both attitudes and behaviours. Future studies can take up generational differences in use of LLMs, and qualitative approaches to understand how human-AI interactions evolve. The implication of the study for parents and educators is the importance of understanding how technology is used and enhancing literacy to promote use without dependence