[目的/意义]探究AI算法诱导“氪时”现象的认知错觉生成机理，厘清其对情报感知与舆论传播的影响，对数字时代信息生态治理具有重要意义。[方法/过程]基于多源异构数据构建多维度分析框架，提出膨胀系数、注意漂移率等量化指标，采用综合建模与案例分析法开展研究，搭建闭环情报感知干预机制。[结果/结论]基于分析结果，从算法优化、用户引导、社群治理三个层面提出认知错觉阻隔策略，旨在为算法优化、用户媒介素养提升及舆情风险治理提供参考。

[Purpose/significance] Exploring the formation mechanism of cognitive illusions induced by AI algorithms in the "time-sink" phenomenon and clarifying its impact on intelligence perception and public opinion communication is of great significance for the governance of the information ecosystem in the digital era. [Method/process] A multi-dimensional analytical framework is constructed based on multi-source heterogeneous data. Quantitative indicators such as expansion coefficient and attention drift rate are proposed. Comprehensive modeling and case analysis are adopted to conduct empirical research, and a closed-loop intervention mechanism for intelligence perception is established.[Result/conclusion] Based on the analysis results, strategies for blocking cognitive illusions are proposed from three dimensions: algorithm optimization, user guidance and community governance, aiming to provide references for algorithm optimization, improvement of users' media literacy and public opinion risk governance.