Artificial Intelligence Assists Mainstream Media in Guiding Public Opinion

The Third Plenary Session of the 20th Central Committee of the CPC emphasized the promotion of integrated management of news propaganda and online public opinion. This is an important direction for the current effort to improve the comprehensive cyberspace governance system. In this context, the concept of "integration" highlights the collaborative cooperation among subjects guiding public opinion, the integration of discourse systems, and the fusion of public opinion fields, requiring mainstream media to carry out news propaganda work with a holistic perspective and systematic thinking. In the context of the New Era, artificial intelligence, as the core driving force of the technological revolution, is profoundly affecting all aspects of politics, economy, and social development.

党的二十届三中全会强调，推进新闻宣传和网络舆论一体化管理。这是当前健全网络综合治理体系的重要方向。在这一背景下，“一体化”理念凸显了舆论引导主体间的协同合作、话语体系的整合以及舆论场域的融合，要求主流媒体以全局视角和系统化思维开展新闻宣传工作。在新时代背景下，人工智能作为科技革命的核心驱动力，正深刻影响着政治、经济和社会发展的方方面面。

The advancement of integrated management of news propaganda and online public opinion guidance is not a smooth path. First, the diversification of public opinion subjects makes the information dissemination environment increasingly open and complex. Internet technology has lowered the threshold for information dissemination, allowing individuals and non-professional institutions to occupy a place in public agenda-setting, with public and private issues intermingling and interacting within the public opinion field. At the same time, while algorithmic recommendation technology improves the efficiency of information distribution, it may also lead to the phenomenon of "information cocoons," limiting exposure to diverse viewpoints and adding difficulty to public opinion guidance. Furthermore, the focus of gatekeeping and regulation has gradually shifted from professional media institutions to internet platforms. Platform media, relying on algorithmic recommendation technology, achieve rapid screening and distribution of massive amounts of information, invisibly controlling the visibility of issues and the direction of topic agendas, further complicating public opinion guidance.

新闻宣传与网络舆论引导一体化管理的推进并非坦途。首先，舆论主体的多元化使信息传播环境日益开放且复杂。互联网技术降低了信息传播门槛，个人和非专业机构得以在公共议程设置中占有一席之地，公共议题与私人议题混杂缠绕，在舆论场中相互激荡。同时，算法推荐技术虽提升了信息分发的效率，却也可能导致“信息茧房”现象的出现，限制多元观点的接触，为舆论引导增添难度。此外，把关监管的重心逐渐从专业媒体机构向互联网平台转移，平台媒体依托算法推荐技术实现对海量信息的快速筛选与分发，无形中控制了议题的可见性和话题议程的走向，进一步加剧了舆论引导的复杂性。

Second, the transformation, misalignment, and emotional characteristics of online discourse are becoming increasingly prominent. The immediacy and fragmented nature of social media platforms allow sudden events to quickly attract public attention and form topic agendas. However, the generation of discourse, the transformation of issues, and the deviation of discussions all exhibit accelerated characteristics, while official and mainstream media responses often face issues of time lag and spatial misalignment. This misalignment is not only temporal—such as the need for sufficient evidence collection and multi-party investigation to ascertain the truth, which takes time, causing official responses to often lag behind the hottest phase of topic discussion—but also spatial, as the dissemination of hot topics and the circle-based spread of diverse viewpoints often isolate mainstream discourse outside groups, making it difficult to engage in effective dialogue with diverse perspectives. Additionally, emotional expression has become a dominant strategy in online discourse; excessive emotional catharsis may exacerbate social division. While undertaking the task of fact-checking, mainstream media must also shoulder the important responsibility of promoting in-depth dialogue and channeling public emotions.

其次，网络话语的转化、错位与情绪性特征日益显著。社交媒体平台的即时性和碎片化传播，使得突发事件能迅速引发公众的关注并形成话题议程。然而，话语的生成、议题的转化以及讨论的偏离都呈现出加速度特征，而官方和主流媒体的回应往往面临时间差和空间错位的问题。这种错位不仅体现在时间上，如调查真相需要进行充分取证和多方调研，耗时较长，官方回应往往滞后于议题讨论最热的阶段；在空间维度上，热点事件的议题散发和多元观点的圈层化传播，使主流话语往往被隔绝在群体之外，难以与多元观点进行有效对话。此外，情感表达成为网络话语的优势策略，过度的情绪宣泄可能加剧社会分裂，主流媒体在承担事实核查任务的同时，还需肩负起促进深入对话、疏导公众情绪的重要职责。

Third, the superposition of cross-media communication and the global public opinion field makes the development of public opinion events more complex and volatile. Cross-media communication, as an important mode of information flow and interaction across different media, has become a routine path for the development of public opinion events. Hot events often first break on a certain platform, then rapidly cross different media, forming widespread social attention and discussion, and generating diverse collisions of viewpoints during the cross-media communication process. However, it is worth noting that the formation of a cross-media public opinion field does not always promote consensus among the public. In this superimposed field, confrontations between different viewpoints and positions become particularly fierce, especially on issues involving national interests, cultural conflicts, and social contradictions, where the antagonistic and conflictual nature of public opinion is more pronounced. How to promptly grasp the process of public opinion development, accurately identify the migration of public opinion fields, and precisely capture the differences in attention between domestic and foreign public opinion has become a prerequisite issue for mainstream media to cope with complex situations and enhance their ability to guide public opinion.

再次，跨媒介传播与全球舆论场的叠加，使得舆情事件的发展更为复杂多变。跨媒介传播作为信息在不同媒介间流布与互动的重要方式，已成为舆情事件发展的常规路径。热点事件往往先在某个平台首次曝光，随后便迅速跨越不同的媒介，形成广泛的社会关注和讨论，并在跨媒介传播过程中形成多元化的观点碰撞。然而，值得注意的是，跨媒介舆论场的形成并不总是能推动公众达成共识。在这一叠加场域中，不同观点和立场的交锋变得尤为激烈，特别是在涉及国家利益、文化冲突和社会矛盾等问题上，舆论的对抗性和冲突性更为明显。如何及时把握舆情发展的过程，准确识别舆论场的迁移，精准捕捉境内外舆论关注的差异，成为主流媒体应对复杂局面、提升舆论引导力的前置问题。

In this context, the widespread application of artificial intelligence technology provides new opportunities for news propaganda and online public opinion guidance. Generative artificial intelligence, with its powerful natural language processing capabilities, can be deeply embedded in the entire process of news production, from content creation to distribution and feedback, comprehensively enhancing the ability of mainstream media to guide online public opinion. The introduction of this technology not only helps mainstream media quickly screen valuable content from massive amounts of information but also, through precise push notifications and personalized services, enhances interaction and stickiness with audiences, thereby more effectively guiding online public opinion in a positive and healthy direction.

在此背景下，人工智能技术的广泛应用为新闻宣传与网络舆论引导提供了新机遇。生成式人工智能凭借强大的自然语言处理能力，能够深度嵌入新闻生产的全流程，从内容创作到分发反馈，全方位提升主流媒体在网络舆论引导方面的能力。这一技术的引入，不仅有助于主流媒体在海量信息中快速筛选出有价值的内容，还能通过精准推送和个性化服务，增强与受众的互动和黏性，从而更有效地引导网络舆论走向积极健康的方向。

Embedding into the scenario construction of news propaganda and public opinion guidance. With the rise of platform media, intelligent algorithms have profoundly reshaped the modes of information dissemination and opinion aggregation, constructing a decentralized communication network. In this context, mainstream media, leveraging their advantages in linking multiple information sources and obtaining accurate and authoritative information, occupy a favorable position in guiding public opinion. With the help of artificial intelligence technology, media organizations can keenly capture public opinion focal points and use information advantages to respond promptly to issues of public concern. Generative artificial intelligence can not only improve the efficiency of content production but also achieve personalized news recommendations by analyzing user behavior and preferences. The combination of authoritative information and precise release gives mainstream media the opportunity to construct a dialogue context and set the discussion tone at the early stage of an event, thereby seizing the initiative in guiding public opinion.

嵌入新闻宣传与舆论引导的场景建构。随着平台媒体的崛起，智能算法深刻重塑了信息传播和观点聚合的模式，构建了一个去中心化的传播网络。在此背景下，主流媒体凭借其链接多方信源、获取信息准确权威的优势，占据了舆论引导的有利地位。借助人工智能技术，媒体机构能够敏锐捕捉舆论焦点，利用信息优势对公众关心的议题进行及时响应。生成式人工智能不仅能够提高内容生产效率，还可以通过分析用户行为和偏好，实现个性化新闻推荐。权威信息与精准发布的结合，使主流媒体有机会在事件初期构建对话语境，设定讨论基调，从而掌握舆论引导的主动权。

Relying on natural language processing and image recognition technologies, media organizations can accurately identify and clean up false information, vulgar content, and rumors, reducing the impact of negative information on mainstream values and creating a clean cyberspace environment. International cooperation and exchange are also important tasks for mainstream media in the context of the New Era. By learning from excellent global media practices and technological innovations, mainstream media can continuously enhance their own communication power and influence. At the same time, strengthening cooperation with international media to jointly address challenges in the global public opinion environment promotes the construction of a more just, transparent, and diverse international information dissemination order. Introducing artificial intelligence technology into the scenario construction of news propaganda and public opinion guidance requires media organizations to transform their information channel advantages into resource aggregation advantages, build platforms rather than fight alone, reintegrate the production process of news reporting, improve the timeliness of reporting, enhance the cohesion of propaganda discourse, and adjust guidance strategies more flexibly.

依托自然语言处理和图像识别技术，媒体机构能够精确识别并清理虚假信息、低俗内容和谣言，降低负面信息对主流价值观的冲击，营造清朗的网络舆论环境。国际合作与交流也是主流媒体在新时代背景下的重要任务。通过借鉴全球优秀的媒体实践和技术创新，主流媒体可以不断提升自身的传播力和影响力。同时，加强与国际媒体的合作，共同应对全球舆论环境中的挑战，推动构建更加公正、透明和多元的国际信息传播秩序。将人工智能技术引入新闻宣传与舆论引导的场景搭建，要求媒体机构能够将自身的信息渠道优势转化为资源聚合优势，搭建平台而非单打独斗，重新整合新闻报道的生产流程，提高报道的时效性，增强宣传话语的凝聚力，更灵活地调整引导策略。

Embedding into the reconstruction system of discourse ecology. In the digital age, text production is no longer limited to competition over content quality but has also engaged in fierce competition around attention and influence. Emotional factors, long overlooked by professionalism, are gradually becoming an important breakthrough for mainstream media in innovating discourse styles and rebuilding discourse ecology. Intelligent technology can rapidly process and analyze massive amounts of real-time data, accurately identify emotional vocabulary, grammatical structures, and emotional evolution in texts, providing support for media organizations to grasp the emotional tendencies and changes of the public and make targeted responses.

嵌入话语生态的重建体系。在数字时代，文本生产已不仅仅局限于内容质量的竞争，更围绕注意力和影响力展开了激烈的角逐。情感要素，这一长期被专业主义所忽视的因素，正逐渐成为主流媒体进行语态革新和重建话语生态的重要突破口。智能技术能够对实时产生的海量数据进行快速处理和分析，精准识别文本中的情感词汇、语法结构以及情感演化，为媒体机构把握民众情感倾向与变化，进行有针对性的回复提供支撑。

Through deep learning on large amounts of emotional data and corpora, artificial intelligence technology can simulate emotionally compelling discourse, helping news media skillfully employ empathy strategies in reporting and propaganda, enhancing the appeal and guiding power of discourse, and making mainstream values deeply rooted in people's hearts. In the future, media can make greater use of AI tools to enhance user participation, establishing closer connections with audiences by creating interactive content and real-time feedback mechanisms. For example, using AI-generated interactive Q&A, public opinion surveys, and topic discussions, media can more systematically understand audience needs and promptly adjust reporting directions to enhance the effectiveness of public opinion guidance. In the field of international communication, generative artificial intelligence can also facilitate cross-language communication, understand and express emotions and concepts in different cultures, generate text content that fits specific cultural backgrounds and contexts, and help mainstream media break down cultural barriers in external communication, enhancing the influence of international communication.

通过对大量情感数据和语料库的深度学习，人工智能技术能够模拟出富有感染力的情感话语，帮助新闻媒体在报道宣传中巧妙运用共情策略，提升话语的感染力和引导力，使主流价值观深入人心。未来的媒体可以更多地利用AI工具增强用户的参与感，通过创建互动式内容和实时反馈机制，与受众建立更紧密的联系。例如，利用AI生成的互动问答、民意调查和话题讨论等，媒体能够更系统地了解受众需求，及时调整报道方向，以增强舆论引导的有效性。在国际传播领域，生成式人工智能还可以进行跨语言沟通，理解和表达不同文化中的情感和观念，生成符合特定文化背景和语境的文本内容，助力主流媒体在对外传播中打破文化藩篱，提升国际传播的影响力。

Embedded throughout the entire process of public opinion identification and guidance. AI can automatically scan diverse information sources, including social media and news websites, collect data in real time, and conduct sentiment analysis, thereby rapidly identifying important issues and potential crises. This capability enables media to intervene in the early stages of an event, effectively disseminating information and guiding public opinion. Through text mining and analysis, media can accurately capture the core issues of public opinion events, providing a scientific basis for news topic selection and planning. After introducing automated news writing, media can quickly respond to public concerns, significantly enhancing the production efficiency of news reporting. The application of generative artificial intelligence is not limited to text content generation but can also handle multiple media forms such as images, videos, and audio. This allows media to adopt multimodal content creation methods and rich information presentation approaches in their reporting, attracting more audience attention.

嵌入舆情识别和舆论引导的全流程。AI能够自动扫描包括社交媒体、新闻网站在内的多元信息源，实时收集数据并进行情感分析，从而快速识别出重要议题和潜在危机。这种能力使媒体能够在事件初期即介入期中，进行有效的信息传播与舆论引导。通过文本挖掘与分析，媒体能够准确捕捉舆情事件的核心议题，为新闻选题和策划提供科学依据。引入自动化新闻写作后，媒体能够迅速回应民众关切，显著提升新闻报道的生产效率。生成式人工智能的应用不仅限于文本内容的生成，还能处理图像、视频和音频等多种媒介形式。这使媒体在报道中能够采用多模态的内容创作方式和丰富的信息呈现方式，吸引更多受众关注。

For example, through AI-generated video summaries and charts, media can more intuitively display complex information, enhancing communication effectiveness. With the help of sentiment recognition and analysis, media can also more accurately grasp the tendencies of issues, effectively alleviating public emotions. In the process of public opinion identification, AI can not only identify hot topics but also predict the evolutionary trends of public opinion. Through comprehensive analysis of historical data and current public opinion, media can anticipate the direction of public opinion in advance, formulate corresponding strategies, and avoid the escalation of public opinion crises. AI technology can also assist media in hierarchical management of public opinion, adopting differentiated handling measures for different levels of public opinion, ensuring the precision and effectiveness of public opinion guidance.

例如，通过AI生成的视频摘要和图表，媒体能够更直观地展示复杂信息，提升传播效果。借助情感识别与分析，媒体还能够更准确地把握议题倾向，有效疏解公众情绪。在舆情识别过程中，AI不仅能够识别热点话题，还能预测舆情的演变趋势。通过对历史数据和当前舆情的综合分析，媒体可以提前预判舆情走向，制定相应的策略，避免舆论危机的升级。AI技术还可以帮助媒体进行舆情分级管理，针对不同级别的舆情采取差异化的处理措施，确保舆情引导的精准性和有效性。

Overall, the application of artificial intelligence technology in the field of news and propaganda provides strong support from a technical perspective for the transformation of media roles, innovation in discourse, and process reengineering. However, we should also recognize that the application of AI technology may also bring potential problems such as content homogenization tendencies and path dependence, constraining media innovation and development. With the popularization of AI technology, media also face challenges in technical ethics and transparency when using generative artificial intelligence. How to maintain the authenticity and credibility of information dissemination while using AI technology is a major issue that mainstream media need to pay close attention to in the future.

总体而言，人工智能技术在新闻宣传领域的应用，从技术层面为媒体的角色转型、语态革新以及流程再造提供了有力支撑。然而，我们也应意识到，人工智能技术的应用同样可能带来内容的同质化倾向和路径依赖等潜在问题，制约媒体的创新与发展。随着AI技术的普及，媒体在利用生成式人工智能时也面临技术伦理和透明度的挑战。如何在使用AI技术的同时保持信息传播的真实性和公信力，是未来主流媒体需要高度重视的问题。

(The author is a researcher at the Research Center for Journalism and Social Development, Renmin University of China)

（作者系中国人民大学新闻与社会发展研究中心研究员）

Online editor: Huihui

网络编辑：慧慧