Editor's note: Currently, generative artificial intelligence is widely used in fields such as information dissemination, public opinion guidance, and knowledge production, and is reconstructing the technological architecture, information dissemination mode, and ideological ecological pattern of cyberspace. In this context, what impact has the development of generative artificial intelligence had on China's ideological construction? How to seize opportunities, mitigate risks, and promote the construction of mainstream ideology? In this regard, Professor Xing Guozhong from the School of Marxism at Beijing Normal University was invited to explore the ways to construct mainstream ideology in the context of technological change. Generative AI empowers efficiency and reshapes the mainstream ideological dissemination pattern. Generative artificial intelligence reshapes the way humans interact with technology and changes the way knowledge and information are produced. What opportunities do you think this will provide for the development of ideological construction? Xing Guozhong: On the one hand, generative artificial intelligence has promoted the reconstruction of content production paradigms. For example, the "Learning Strong Country" official document assistant, jointly developed by the "Learning Strong Country" learning platform and Baidu's generative model Wen Xiaoyan, integrates functions such as document retrieval, intelligent questioning, and information extraction, meeting users' efficient needs for official document information and reshaping the path of mainstream ideological dissemination in the digital age. Generative artificial intelligence not only achieves effective integration of public and private domain resources and collaboration of knowledge information, but also expands knowledge boundaries based on big model data training, effectively expands the scope of ideological content dissemination, and achieves cross media and cross platform content dissemination. On the other hand, generative artificial intelligence can help improve the efficiency of precise communication. The continuous interaction between generative artificial intelligence and humans can generate ideological content in real-time that meets user preferences and habits. Users can issue instructions to the intelligent system through natural language, obtain outputs that match their needs and values, thereby enhancing macro perception, producing a large amount of content that conforms to mainstream ideology, and enhancing the effectiveness of ideological guidance. For example, the first AI animated series in China, "Ode to the Poetry of a Thousand Years," utilized a large number of traditional Chinese paintings and ancient poems as training materials, communicated with generative models, and created AI animations with unique Chinese aesthetics, innovating the expression of mainstream values and enhancing the appeal, persuasiveness, and guidance of socialist ideology. What are the new features or challenges that mainstream ideological construction faces in the context of the multidimensional challenges brought about by the application of technology? How do these characteristics change the operational logic of traditional ideological work? Xing Guozhong: Generative artificial intelligence combines technical and value attributes, and its application presents a more complex environment for ideological construction, presenting the following new features: firstly, generative artificial intelligence has triggered a paradigm shift towards decentralized information production. Generative artificial intelligence endows users with almost zero threshold content creation capabilities, breaks the monopoly of traditional professional institutions on information production, broadens channels for individuals to express their opinions, and diversifies forms of expression. The game field of ideological discourse power has shifted from the traditional binary structure of "institution audience" to a complex network ecology of "algorithm user platform", where each node may become a trigger point for ideological dissemination. Secondly, generative artificial intelligence constructs a cognitive construction space that blends reality and virtuality. The boundary between virtual content generated by generative artificial intelligence and real information is blurred, and false information may become an ideology with potential action guidance at the value orientation level, and even evolve into public rumors. When false information gains super real dissemination power, the "truth verification mechanism" of ideology becomes ineffective, and public cognition forms a nested structure of "re mimicry in a mimetic environment", which may consume social trust, dilute or even disintegrate mainstream ideology. Again, generative artificial intelligence can achieve fine-grained cognitive intervention. Through user profiling and natural language processing technology, generative artificial intelligence can achieve precise delivery of ideological dissemination, form "customized narratives" targeting the cognitive characteristics of different groups, strengthen the inherent cognition of different groups or change their cognitive structure, unconsciously establish users' ideology, and provide highly persuasive experiences in a clear and organized manner, reducing users' questioning and criticism of other information. Finally, generative artificial intelligence may lead to a crisis of identity in subject reconstruction. Generative artificial intelligence has demonstrated human like abilities such as generating meaningful information, precise push, and reviewing and processing information, which to some extent replace some of human labor. However, it also poses a risk of depriving labor as a fundamental human ability and impacting human subjectivity. At the same time, the ideological content generated by generative artificial intelligence often hides algorithmic biases and technological ethical dilemmas. When users cannot distinguish the human intentions and machine logic behind the content, the trust foundation of mainstream ideology faces the risk of deconstruction. Adhere to the leadership of the Party and innovate governance simultaneously, and strengthen the mainstream ideological position in the intelligent era. Based on the new changes you mentioned, how do you think China should balance opportunities and risks, and strengthen the construction of mainstream ideology? Xing Guozhong: In the context of the increasing depth of intelligent algorithm technology in information dissemination, the dissemination strategy of socialist ideology needs to actively adapt to and integrate into this change. By cleverly integrating new algorithm technology into the mainstream ideological dissemination framework in a flexible way, innovation and upgrading of dissemination methods can be achieved. Firstly, adhere to the leadership of the Party. We should adhere to the principle of the Party managing data, integrate Party media data and socialist ideology content into generative artificial intelligence databases, create data resources with "warmth", and make generative artificial intelligence products a "new mouthpiece" for spreading mainstream ideology. Strengthen the technological empowerment and people-oriented concept in the smart ecosystem, ensure the people-oriented nature of data, and deeply explore the value of data that truly reflects the voice of the people. Strengthen the supervision of providers of generative artificial intelligence big models and establish an ideological work responsibility system with collaborative participation of all parties under the leadership of the Party. In short, it is necessary to ensure that the Party occupies an active and advantageous position in the development of intelligent technology, and that the mainstream ideology occupies a dominant position in the market of generative artificial intelligence. Secondly, improve the legal and regulatory system. The current laws need to be improved in terms of the obligation of value transmission among various entities, and legal supervision needs to be strengthened. It is necessary to clarify the boundaries of the use of generative artificial intelligence through the national legal system, delineate ideological management red lines, and attach great importance to the legal supervision of data collection, intelligent algorithm operation, and generated content for the training of generative artificial intelligence big models. We should establish a generative artificial intelligence data supervision framework, improve the data risk complaint and reporting mechanism, refine risk prevention measures, and build a unified data risk responsibility system. Establish algorithm standards for intelligent technology applications, improve algorithm filing and review procedures, explore public interest litigation channels for algorithm infringement, standardize personalized recommendation algorithms, and prevent algorithm risks. Pay attention to the intellectual property and legal responsibilities of generated content, protect the copyright of creators, clarify the responsibilities and obligations of all parties, further improve the definition of artificial intelligence identity and its social behavior legal norms, and ensure that the legal system is compatible with technological development. Thirdly, adhere to technological innovation. Firstly, promote technological innovation through the layout of the entire industry chain. Deepen national strategic planning, increase research and development investment in key areas such as data, algorithms, computing power, and applications, promote the deep integration of the entire industry chain of generative artificial intelligence with the primary, secondary, and tertiary industries, and achieve the organic combination of technology and ideological governance, in order to synchronize technological development with national strategic goals. Secondly, by regulating capital development, we can help promote technological innovation breakthroughs. The government needs to promote the deep integration of technology enterprises with mainstream ideological dissemination through capital guidance and policy regulation, and prevent short-sighted behavior of capital. Finally, we need to strengthen talent cultivation and solidify the foundation of technological innovation. Improve the talent cultivation mechanism, closely align with the needs of the industry chain, strengthen the talent cultivation of key positions such as data architects, algorithm engineers, data engineers, and artificial intelligence trainers, and provide solid talent support for technological development. Fourthly, innovative discourse expression. Currently, narrative issues based on generative artificial intelligence often manifest as fragmented expression and pan entertainment, leading to the gradual dilution of serious and positive mainstream ideology in humorous entertainment. Therefore, mainstream media should enhance their ability to set issues, by creating narrative topics that are close to people's lives, such as social governance, technological innovation, and green development. They should build an issue communication environment that meets the personalized needs of the audience and enhance the leading power of mainstream ideology. Mainstream media should enhance the effectiveness of ideological narrative. We should fully leverage the advantages of artificial intelligence technology, analyze user preferences, utilize diverse narrative modes, generate rich narrative styles, use narrative techniques such as suspense and contrast, achieve three-dimensional dissemination and sensory expression of ideological discourse, update content presentation forms through intelligent means, and enhance value recognition. Fifth, enhance intelligence literacy. Consciously and continuously improving users' intelligence literacy helps optimize the human-computer interaction information ecosystem, correct the biased ideological values embedded in generative artificial intelligence, and strengthen human subjectivity. To cultivate users' awareness of data subjects, enhance their proactive awareness of data collection, processing, utilization, and transformation, and form a correct data concept that adapts to the intelligent era. To enhance users' data expression literacy, allowing them to dominate the feedback loop of human-computer interaction through the dissemination of a large amount of positive information, effectively overcoming the prior ideological bias that language models may carry. To enhance the user's ability to use generative artificial intelligence, the country should make full use of publicity and education resources, comprehensively promote intelligent technology knowledge, cultivate the public's scientific understanding of intelligent technology, and enable them to accept the changes brought by generative artificial intelligence with a rational attitude. (New Society)