Empowerment or disempowerment? How generative AI consultation shapes the health decision-making among the new generation of older adults.

Generative artificial intelligence has lowered the threshold for access to health information, but increased access does not equate to optimal use. While the new generation of older adults benefit from greater convenience, their decision-making risks have not decreased correspondingly, thereby constituting a new health decision-making paradox. To uncover the underlying mechanism by which generative AI health consultations influence decision quality, this study integrates triadic reciprocal determinism and metacognitive theory, constructing an analytical framework with "generative AI health consultation-intelligent information-health decision quality" as the core pathway. Intelligent information is divided into five stages: attention, perception, discernment, trust, and reliance. This study introduces intelligent boundary awareness, algorithmic pandering awareness, and health autonomy orientation as moderating variables, thereby characterizing differentiated effects under varying cognitive and behavioral contexts. Furthermore, this study employs a hybrid methodological approach that combines structural equation modeling and fuzzy-set qualitative comparative analysis to test the model, identifying three types of high-quality health decision-making patterns: trust-reinforced, autonomous adaptive, and cognitive adoption modes. The results indicate that: (1) Generative AI health consultation improves decision quality not merely by increasing the supply of information, but by reshaping new generation of older adults' cognition and engagement patterns with intelligent information; (2) Intelligent boundary awareness, algorithmic pandering awareness, and health autonomy orientation significantly moderate the translation of intelligent information into decision quality, thereby revealing cognitive and agentic boundaries of decision efficacy; (3) The three identified patterns transcend linear explanations based on single factors, health decision quality depends on differentiated alignment among information cognition, technological relationships, and individual autonomy. Therefore, the key to enhancing the efficacy of generative AI health consultations is to shift from tool provision to empowerment, with a primary focus on strengthening intelligent information discernment, rational trust, and autonomous decision-making capabilities in the new generation of older adults.