Multi-layered Epistemic Disruption in AI-Driven Health Misinformation: Conceptual Framework and Viewpoint (Preprint)

BACKGROUND: Generative AI (GenAI) has transformed the health information ecosystem by enabling scalable, sophisticated health misinformation production at near-zero marginal cost. Current literature addresses AI's role in health misinformation predominantly through a binary threat-detection framework, systematically overlooking the structural, multilayered mechanisms through which AI simultaneously embeds false claims across intersecting human trust systems. OBJECTIVE: This paper introduces the Multi-layered Epistemic Disruption Framework (MEDF), which conceptualizes how AI-driven health misinformation structurally undermines public trust through four interdependent layers of cognitive and institutional disruption: discursive (clinical language shielding: fluent medical terminology and fabricated citations deployed as credibility signals), biometric (embodied authority transfer: deepfake appropriation of real clinicians' faces and voices), temporal (the synthetic chorus effect: near-simultaneous fabrication of apparently independent corroborating sources), and systemic (structural epistemic erosion: cumulative macro-level collapse of trust in medical institutions). METHODS: Adopting a socioecological and structural epistemic approach, this Viewpoint synthesizes empirical findings from communication psychology, medical sociology, and digital infodemiology. The MEDF is explicitly positioned relative to established health communication frameworks, including the i-frame/s-frame distinction (individual-level vs system-level intervention targets) and socioecological infodemic models, and each construct's novelty is defined in relation to adjacent concepts in prior literature. RESULTS: The MEDF proposes that AI-driven health misinformation is distinctively dangerous due to its capacity to exploit variable individual receptivity to medical authority claims and to simultaneously lower epistemic thresholds across multiple trust layers. Population-level data indicate that individuals who frequently encounter health misinformation on social media are 1.66 times more likely to report systemic distrust of healthcare institutions (OR 1.66; 95% CI 1.11-2.48). Perceptual studies document that listeners correctly identify AI-generated voice clones only about 60% of the time and perceive a cloned voice as identical to its real counterpart in approximately 80% of trials. Existing defenses - including C2PA provenance standards, automated deepfake detection (showing AUC drops of up to 50% under real-world conditions), and prebunking interventions - are shown to address only subsets of the proposed cascade, leaving temporal and systemic layers substantially unmitigated. Four testable hypotheses are advanced for empirical validation. CONCLUSIONS: Addressing AI-driven health misinformation requires moving beyond individual-level i-frame interventions toward structural, s-frame policy responses calibrated to each layer of the MEDF cascade. Policymakers and platforms must implement source identity verification, clinician biometric protection protocols, cross-platform ecosystem governance, and proactive trust infrastructure, with particular urgency in lower- and middle-income country (LMIC) contexts where regulatory capacity and platform oversight are most limited.