Document Type

Article

Date of Original Version

9-24-2026

Abstract

Generative AI, encompassing large language models, image generators, and voice synthesis tools, has transformed the information landscape in ways that extend well beyond the production of fabricated content. By undermining the evidentiary signals through which individuals and communities have historically calibrated epistemic trust, it poses a structural challenge to existing media literacy frameworks. These frameworks, built around a “detect-and-debunk” paradigm, equip learners to identify false content through source verification, fact-checking, and evidence evaluation. These skills remain valuable but rest on assumptions that generative AI has rendered progressively less tenable. This article proposes “epistemic resilience” as a complementary goal: the cultivated capacity of individuals and communities to sustain sound epistemic practices under conditions of pervasive informational uncertainty and institutional erosion. Drawing on a theoretically grounded conceptual synthesis of scholarship in social epistemology, trust theory, cognitive psychology, and digital media studies, it develops a four-dimensional framework, comprising calibrated trust, evidential reasoning under uncertainty, metacognitive awareness, and participation in epistemic communities, and evaluates existing assessment instruments, including the Digital Online Media Literacy Assessment (DOMLA), against this framework. The article identifies specific modifications to existing tools, examines the policy and pedagogical implications of the framework, and addresses the implementation challenges these imply. It concludes by proposing a research agenda to move the framework from a conceptual proposal to empirically grounded practice.

Download

Share

COinS