From “Information Cocoons” to “Cognitive Hallucination”: The Dissemination Mechanisms and Governance Dilemmas of Health Information Disorder in Algorithmic Recommendation Environments

Objective: This study examines how algorithmic recommendation environments reshape the dissemination mechanisms of health information disorder and explores why existing governance strategies have become increasingly ineffective.

Methods: Drawing on the perspectives of philosophy of technology and the political economy of communication, the paper adopts a theoretical and literature-based analytical approach to construct a three-dimensional framework of technology, commerce, and cognition.

Results: The study shows that health information disorder in algorithmic environments is generated through three interrelated mechanisms: probabilistic hallucination in large-model generation, systematic attention exploitation driven by platform logic, and multimodal hallucinatory effects that weaken users’ critical judgment, especially among older adults. These changes indicate that the traditional concept of the “information cocoon” is no longer sufficient to explain the complexity of current health information risks. Instead, a new form of risk, termed cognitive hallucination, has emerged. The findings further reveal that governance approaches centered on content blocking face temporal, spatial, and structural limitations.

Conclusion: Effective governance should move beyond ex post deletion and toward front-end intervention. The concept of cognitive friction provides a useful theoretical basis for redesigning governance strategies and advancing age-friendly digital governance in algorithm-mediated health communication.