Grade-Level Differences in Adolescent Anthropomorphism Toward Conversational AI



Abstract Book of the 10th International Conference on New Trends in Social Sciences

Year: 2026

[PDF]

Grade-Level Differences in Adolescent Anthropomorphism Toward Conversational AI

Raunak Garhyan

ABSTRACT:

Adolescent interaction with conversational artificial intelligence (CAIs), most notably large language model (LLM) based systems, has led to critical questions about anthropomorphic interaction and its consequences to human-computer interaction and education. Learning, comforting, sharing, talking to LLMs may create a sense of connection. In this research, high school students were surveyed to investigate five anthropomorphic dimensions viz roleplay, courtesy, emotionality, reinforcement, and companionship. The interaction of these anthropomorphic dimensions with perceived functionality of the CAs was studied with respect to gender and grade level on a sample of 109 adolescents. Findings showed that there were no significant gender differences in dimensions of anthropomorphism or perceived functionality. However, the differences in grades levels were significant, underclassmen (Grades 9-10) reported an increase in all dimensions and perceived functionality, with large effect sizes. Selective effects were observed in interaction time with CA, and these differences were significantly observed in roleplay. Correlational and regression analysis indicated that some of the anthropomorphic dimensions, especially courtesy, were positively related to and predictive of the perceived functionality but the relationship between them varied by grade levels. Altogether, the current results indicate that anthropomorphic interaction with CAs is not uniform across adolescence and is associated with how students evaluate the functional effectiveness of these systems. The study is among the first to examine adolescents across all five anthropomorphic dimensions, and its findings may inform the development of guardrails for educational technologies as well as age-sensitive AI literacy frameworks.

Keywords: Conversational AI, Anthropomorphism, Artificial Intelligence, Perceived Functionality, Companionship, Emotionality, Human-Computer Interaction, Educational Technology