【目的/意义】聚焦大语言模型在事实核查领域的应用效果，系统探究其对用户信息真伪甄别效果的影响，并考察人口统计学变量的调节效应，为社交平台打击虚假信息提供实证支持。【方法/过程】以健康类信息为具体情境和实验材料，基于双过程理论采用组间对照实验设计，招募142名参与者，构建使用LLM进行事实核查的实验组和依靠自身知识判断的对照组，通过单因素方差分析与调节效应分析，探究LLM对用户信息真伪甄别效果的影响及性别、年龄、受教育程度的调节作用。【结果/结论】LLM辅助事实核查显著提升用户信息的真伪甄别效果，其识别虚假信息准确率显著提高，但对真实信息的识别准确率提升未达到显著水平。进一步分析表明，低受教育程度用户在借助LLM辅助事实核查时，信息真伪甄别效果提升更为突出，而性别与年龄变量未呈现显著调节效果。

[Purpose/Significance] This study focuses on the application effect of large language models in the field of fact-checking, aiming to systematically explore their impact on users' ability to distinguish the authenticity of information and examine the moderating effect of demographic variables, providing empirical support for social platforms to combat false information. [Method/Process] This study used health-related information as the specific context and experimental material, adopting an inter-group control experiment design based on the dual-process theory. A total of 142 participants were recruited, and an experimental group using LLM for fact-checking and a control group relying on their own knowledge for judgment were established. Through one-way ANOVA and moderation effect analysis, the impact of LLM on users' ability to distinguish the authenticity of information and the moderating effects of gender, age, and education level were explored. [Result/Conclusion] LLM-assisted fact-checking significantly improves users' ability to distinguish the authenticity of information, with a significant increase in the accuracy of identifying false information, but the improvement in the accuracy of identifying true information does not reach a significant level. Further analysis indicates that users with lower educational attainment show a more pronounced improvement in their ability to distinguish the authenticity of information when assisted by LLMs in fact-checking, while gender and age variables do not exhibit significant moderating effects.