Effects of anthropomorphic features of generative artificial intelligence tools on learners’ role perceptions: evidence from interview and fNIRS

View at publisher Summarise with VIRAT
Authors: Liangliang Xia, Dong Yi, Jessie Siew-Pin Leuk, Wei‐Peng Teo
Cited by: 0
References: 48

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

While generative artificial intelligence (GenAI) is increasingly being designed for human-like roles to facilitate learning, the extent to which learners perceive GenAI as a real human remains unclear. This study investigated the impact of anthropomorphic features of GenAI on learners' role perceptions and their neural correlates. A 2 (task difficulty: low and high) ×3 (anthropomorphic features: control, low, and high) mixed design was conducted. 121 participants were assigned to the control group (CG), low-anthropomorphic GenAI group (LAG), and high-anthropomorphic GenAI group (HAG). The interview was employed to investigate subjective perceptions, while functional near-infrared spectroscopy (fNIRS) was used to measure neural activation of the prefrontal cortex (PFC) and right temporoparietal junction (r-TPJ). Interview results showed that 86% of HAG perceived GenAI as human-like roles, compared to 13.2% in LAG and 0% in CG. Moreover, the main effects of task difficulty and anthropomorphic features on neural activation were significant. The PFC activation was higher in the difficulty task than in the simple task. Neural activation in sub-regions of the r-TPJ was higher in HAG than in CG and LAG. These findings provide empirical insights into the potential for GenAI to fulfill human-like roles and illuminate the neural mechanisms underlying role perceptions.

Fields of Study & Subjects

Social Robot Interaction and HRI Intelligent Tutoring Systems and Adaptive Learning Artificial Intelligence in Education

Publication details

Journal / source
Interactive Learning Environments
Volume
34
Issue
8
Pages
5948–5963
Publication date
27 March 2026
Language
EN