Long-term effects of LLM-based AI tutors on child and adolescent development

Determine how prolonged interaction with AI tutors powered by large language models affects psychological development in children and adolescents, specifically whether such interactions reinforce or challenge cognitive biases and whether they cultivate a persistent tendency to outsource critical thinking to AI systems.

Background

The paper argues that LLMs are increasingly embedded in educational contexts as AI tutors, yet their underlying value expressions and potential developmental impacts remain unassessed by standard research tools. The authors emphasize that psychometric assessment, enabled by their Ethics Engine pipeline, can measure value expression but that the long-term effects on developing minds are still unknown.

Within the broader theme of making LLM assessment accessible to non-technical researchers, the authors highlight that the pedagogical values and sustained influence of AI tutors on cognition and critical thinking have not been empirically characterized. This motivates targeted longitudinal and developmental research using accessible measurement infrastructure.