JMIR Publications

ABSTRACT

The role of artificial intelligence in health misinformation is predominantly framed in the literature as a dual-role threat-opportunity dichotomy. This paper argues that this framing is insufficient and proposes a more comprehensive analytical model specific to health information ecosystems: the Multi-layered Epistemic Disruption Framework (MEDF). The framework contends that the danger of AI lies not in generating false information per se, but in its capacity to embed false information simultaneously across four interconnected layers of human trust systems. Although these layers are conceptually distinct, in practice they are intertwined and produce a cumulative effect: (1) the discursive layer — clinical language shielding; (2) the biometric layer — embodied authority transfer; (3) the temporal layer — the synthetic chorus effect; (4) the systemic layer — structural epistemic erosion. This paper is presented as a perspective/framework piece; while the model has not been empirically tested, testable hypotheses are developed for each layer and research priorities are identified.