Abstract

AI-enabled learning technologies are rapidly becoming institutional infrastructure. In this context, “trust” cannot be treated as a simple user attitude or a byproduct of technical performance. We argue that durable trust in AI-enabled learning systems is fundamentally a legitimacy outcome: stakeholders must judge AI-mediated practices as appropriate, credible, and defensible within the normative and governance structures of learning environments. Building on socio-technical systems scholarship and institutional legitimacy theory, we introduce Socially Legitimized Trust (SLT), a framework that helps us understand adoption stability and contestation through alignment among three foundations: technocratic validation (credible evidence of capability, limits, and reliability), social validation (shared normative judgments among instructional actors regarding appropriateness and credibility), and institutional authorization (formal governance, policy, and accountability mechanisms that allocate decision rights and responsibility). Further, we argue that SLT predicts recognizable instability patterns under misalignment, including contested use despite high technical performance, fragile informal practices under weak authorization, and compliance without confidence when authorization outpaces validation. We formalize five propositions to guide empirical research on stability, contestation, cross-boundary credential credibility, and design orientations that favor learning augmentation over task substitution. By shifting the unit of analysis from individual users or artifacts to the learning system, SLT provides an integrative agenda for studying how trust is constructed, challenged, and repaired as AI becomes embedded in learning infrastructures.

Description

For all papers published in AIRCC journals, the copyright of the paper is retained by the author under Creative Commons (CC) Attribution license. This license authorizes unrestricted circulation and reproduction of the publication by anybody, as long as the original work is properly cited.

Publisher

AIRCC

Date of publication

Spring 3-2026

Language

english

Persistent identifier

http://hdl.handle.net/10950/5078

Document Type

Article

Publisher Citation

Patole, S. R., Carpenter, R. E., Milewicz, R. M., Baker, R. M., & Turner, J. R. (2026). Towards socially legitimized trust in AI-learning systems. International Journal on Integrating Technology in Education, 15(1), 17–36. https://doi.org/10.5121/ijite.2026.15102

Download

Share

COinS