Perceived Anthropomorphism as a Mediating Mechanism in Human–AI Interaction: Explaining Empathy, Trust, and Perceived Value in Higher Education
Artificial Intelligence (AI) is increasingly evaluated as both a functional technology and a socially responsive interaction partner. This study examines how academic staff form trust in educational AI and how that trust contributes to perceived value. Its central contribution is to position perceived anthropomorphism as a mediating mechanism through which perceived animacy and perceived intelligence are translated into perceived empathy. A cross-sectional survey was completed by 236 academic staff members, and the model was tested using partial least squares structural equation modeling. The measurement model demonstrated satisfactory reliability and validity. A competing-path analysis provided a more differentiated account of the proposed mediation mechanism. In the final competing-path model, Perceived Usefulness remained the strongest predictor of Perceived Value (β = 0.586), while Perceived Intelligence strongly predicted Perceived Usefulness (β = 0.570) and also had a substantial direct effect on Perceived Empathy (β = 0.465). Perceived Anthropomorphism remained positively associated with Perceived Empathy (β = 0.238). The direct effect of Perceived Animacy on Perceived Empathy was not significant (β = 0.104, p = 0.154), whereas its indirect effect through Perceived Anthropomorphism remained significant (β = 0.095, p = 0.006), indicating indirect-only mediation. For Perceived Intelligence, both the indirect effect through Perceived Anthropomorphism (β = 0.057, p = 0.020) and the direct effect on Perceived Empathy were significant, indicating complementary partial mediation. Perceived Usefulness and Trust in AI jointly explained 57.3% of the variance in Perceived Value (R2 = 0.573). The findings reveal two complementary routes to valuing educational AI: a functional–cognitive route centered on intelligence and usefulness, and a social–emotional route through which anthropomorphism converts AI features into perceived empathy.
Authors
- Mohammed Al-Mamari (ORCID: https://orcid.org/0000-0001-5772-9888)
- Abdullah Al-Abri (ORCID: https://orcid.org/0000-0001-7198-9643)
Institutions
- Rustaq College of Education (OM)
- College of Applied Sciences, Nizwa (OM)
Publication Details
- Journal
- Education Sciences
- Published
- 2026-09-20
- DOI
- https://doi.org/10.3390/educsci16091558
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 0.00