Cognitive fog in AI-assisted care: Preserving clinical reasoning, patient education, and shared decision-making

This discussion paper conceptualizes cognitive fog as a communication and reasoning problem that can emerge when clinicians, patients, or organizations over-rely on artificial intelligence (AI) tools in health care.

Drawing on a selective, theory-oriented narrative synthesis of literature on AI-enabled clinical decision support, large language models, automation bias, cognitive offloading, patient-facing AI communication, and shared decision-making, this paper defines cognitive fog as a condition in which fluent, rapid, and institutionally embedded AI output blurs the boundary between assistance and authority. It comprises epistemic blurring, metacognitive weakening, and relational displacement. In clinical and patient-facing contexts, inaccurate or biased recommendations may reduce accuracy, while polished language can make generic education appear personal and automated recommendations appear deliberative. Management requires explicit role boundaries, AI literacy, cognitive forcing routines, visible human review, proportionate patient-facing disclosure, and governance that addresses workflow, accountability, equity, and patient understanding.

AI overreliance is not only a technical safety issue; it is also a health communication issue because it can cloud reasoning, weaken accountability, and narrow shared decision-making. Effective mitigation must address both individual barriers, including time pressure, variable AI literacy, and reluctance to challenge automated output, and systemic barriers, including weak governance, opaque tools, misaligned incentives, and inadequate monitoring.