Core Concept - Tencent Research Institute introduces the concept of "Information Hive" to counter the "Information Cocoon" phenomenon, emphasizing active user participation in a collaborative information ecosystem [1][20]. Group 1: Characteristics of Information Hive - The "Information Hive" promotes diverse information sources, allowing users to navigate multiple channels rather than being limited to algorithmic recommendations [3][13]. - It encourages strong user agency, enabling users to actively explore and select information rather than passively consuming content [4][5]. - Collaborative construction is emphasized, where users not only consume but also create, disseminate, and evaluate information, shaping the ecosystem collectively [6][7]. Group 2: Design and Functionality of Information Products - Effective information products should facilitate user-generated content, ensuring diverse perspectives are included in the ecosystem [8][10]. - The design should promote interconnectivity among different "hives," allowing free flow of information rather than being restricted by algorithms [9][10]. - Technical measures such as open APIs, cross-platform tools, and standardized content formats are essential for enhancing information accessibility and interaction [10][11]. Group 3: User Behavior and Engagement - Users should be guided towards proactive exploration and utilizing multiple entry points for information, avoiding reliance on a single information stream [12][21]. - Encouraging community collaboration and feedback mechanisms can enhance user engagement and knowledge sharing across different platforms [11][12]. - The focus should be on creating a dynamic, responsive information ecosystem that adapts to user needs and promotes long-term knowledge accumulation [20][22]. Group 4: Examples of Information Hive Products - Wikipedia exemplifies an "Information Hive" product with its open collaborative editing model, allowing users to contribute and access knowledge freely [13]. - Quora and Douban serve as platforms that encourage active user participation and diverse knowledge sharing through community interactions [14][15]. - Reddit functions as a network of diverse communities, enabling users to engage in discussions and share information across various topics [15][16]. Group 5: Future Directions - The concept of "Information Hive" is still in its early stages and requires further empirical research to solidify its theoretical foundation and practical applicability [20]. - The optimization of information ecosystems should focus on enhancing user agency, promoting diversity in content, and fostering collaborative knowledge creation [21][22].

生成式人工智能生成的内容将"污染"互联网，这是自从ChatGPT席卷全球之后，许多有识之士预言的未 来。现实也确实如此，从知乎、小红书、抖音等内容平台到微信朋友圈，再到淘宝、拼多多，AI生成 的垃圾内容几乎渗透到了互联网的每个角落。 随着AI生成的内容越来越多，互联网上真正"人写的东西"已经变得越来越难找。为此，维基百科方面近 日做出了一项新的决定，赋予管理员更高的权限，在满足特定条件时可迅速删除AI生成的内容。 用维基百科编辑的话来说，"总体而言，AI 内容的泛滥被描述为对维基百科的一种'生存威胁'，我们的 工作流程向来依赖（往往是冗长的）讨论与共识构建，若没有相应的快速删除机制，AI可以极高的效 率地生成大量虚假内容，就将成为严重的问题。" 具体来说，维基百科将立刻删除"AI味过重"的内容，比如"作为一个大语言模型，我XXXX"等生成式AI 工具的经典话术，或是引用出现明显错误的内容，这也是生成式AI最典型的错误，它们会引用根本不 存在的文献、论文。 其实维基百科的内容创作者和内容消费者一致抵制AI是有理由的，毕竟作为全球最大、且最受大众欢 迎的网络百科全书，维基百科是由全球各地的志愿者合作编撰而成，它的 ...