Luo Yuting and Xiao Tianle: Innovation of Online Ideological and Political Education from the Perspective of Algorithmic Recommendation
Xi Jinping pointed out at the National Conference on Ideological and Political Work in Universities and Colleges: "We must utilize new media and new technologies to make our work come alive, promoting the high-level integration of the traditional advantages of ideological and political work with information technology, and enhancing its sense of the times and its appeal." This important discourse profoundly reveals the vital role of new technologies in strengthening and improving cyber ideological and political education. Currently, as a significant form of artificial intelligence technology, algorithmic recommendation is widely applied across numerous network platforms and has already become the primary mode of online information distribution. Ideological and political educators must be adept at utilizing this new technology of algorithmic recommendation to further enhance the sense of the times and the precision of cyber ideological and political education, improving its appeal and guiding power.
I. Algorithmic Recommendation: A Form of Discourse Power Based on Big Data Technology
Algorithmic recommendation is a form of discourse power derived from big data technology. It can both precisely identify people's needs and precisely satisfy them, continuously elevating its discourse power through the process of precision identification and distribution. In a certain sense, whoever possesses algorithmic technology possesses the discourse power of the algorithmic era.
1. Algorithmic recommendation is a technology: the "art of mind-reading"
With the rapid development of internet technology, humanity has entered an era of information explosion. Faced with an ocean of network information, quickly locating the information one needs has become a core requirement for network users. Consequently, algorithmic recommendation technology, characterized by precision delivery and effective supply, emerged as the times required. Algorithmic recommendation utilizes big data technology to track users' online behavior, thereby calculating their interests and preferences. On the basis of "understanding you," it delivers information of interest with "one-on-one" precision. To a certain extent, algorithmic recommendation has solved the difficult problem of "matching massive information with individual needs" and is able to precisely grasp and satisfy individual psychological needs. Therefore, some have figuratively called it the "art of mind-reading," and it has quickly developed into the primary method of current online information distribution. The technical mechanism of algorithmic recommendation mainly includes the following three steps:
First, "information identification," which involves the tagging of network information. Intelligent algorithms perform standardized and data-driven processing of massive information through tagging and categorical marking to achieve efficient screening and precise identification. During processing, based on different information types such as text, images, and video, the algorithm decomposes original information into character strings and bytes [1] that the machine can identify and use to calculate similarities and differences. That is, it converts raw information content that the machine cannot recognize into identifiable data features and adds corresponding tags to these features. Tags are the most important metadata of information; they describe the attributes of the data and reveal the relationships between data points. Through information tagging, data management can be carried out effectively, allowing for the rapid identification and retrieval of required information within massive datasets.
Second, "user profiling," which involves the modeling of algorithm users. Generally speaking, a user's interests and preferences remain relatively stable over a period of time. Therefore, based on the extensive collection of user information, a user interest preference model can be established and analyzed to predict future behavior. "User profiling" presents a user's behavior and interests in the form of data, transforming user information into data and conducting model analysis on this basis. The accuracy of the "user profile" depends on whether the user information is comprehensive; thus, user information must be collected extensively. User information mainly includes: personal attribute information (gender, age, region, ethnicity, education, occupation, etc.), historical behavior information (browsing traces, browsing duration, browsing frequency, etc.), and behavioral context information (access time, access location, etc.). The algorithm will perform multi-dimensional profiling of the user based on the synthesis of this information and data. It should be noted that "user profiling" is not static; the algorithm will continuously update the model based on user feedback, using dynamic adjustments to improve the accuracy of pushed information.
Third, "precision delivery," which involves adaptive processing for algorithm users. Based on the degree of matching between information tags and user profiles, algorithmic adaptation is performed between the information and the user, forming a recommendation list and ranking it to deliver information precisely. In the "precision delivery" phase, tagging remains the key factor. Simply put, tags represent the types of user interests, while the frequency of tag usage reflects the degree of the user's preference for certain types of information. Based on obtaining tag types and usage frequency, and combining this with user feedback and specific usage contexts, the algorithm comprehensively applies different algorithmic model strategies to form the final tailor-made information content and push it precisely to specific users.
2. Algorithmic recommendation is a power: discourse power
"Power means every opportunity within a social relationship to carry out one's own will even against resistance, regardless of the basis on which this opportunity rests." [2] Accordingly, we can understand power as a relationship of dominance and constraint exerted by a subject of power over an object of power due to the possession of a specific superior position. As a force for conquering and transforming nature, technology itself has no value loading and does not possess a dimension or meaning of power. However, if technology breaches the boundary of "man's dominance over things" and involves relations between people, thereby impacting people's real interests, then technology acquires the attribute of power.
Algorithmic recommendation influences people's thoughts and even drives their actions imperceptibly through means such as information filtering, agenda-setting, and discourse communication. Therefore, algorithmic recommendation belongs to a type of hidden discourse power. First, algorithms utilize information filtering to shape people's thoughts through the dissemination and acquisition of dominant values. Although the algorithm itself may have no bias toward information content, the algorithm designer can regulate the content and quantity of information through technical means, thereby gaining the power to filter and select information. In the process of designing algorithmic models according to their own goals and intentions, algorithm designers largely determine which information will ultimately be recommended and presented to users; the value preferences of the designer determine the content and quantity of value-laden information the user can obtain. Second, algorithms use agenda-setting and discourse communication to influence people's thoughts by controlling the direction of public opinion. Cyberspace is a space of "polyphonic clamor" [3]; the influence of an individual agenda is extremely limited, but when numerous netizens focus on a single topic simultaneously, this influence is staggering. Algorithmic recommendation can "set the pace" [4] by densely pushing the same type of discourse information in a short period, continuously strengthening the value identity and dissemination within a specific group. This quickly transforms minor individual agendas into hot public agendas and small-circle discourse communication into large-scale discourse dissemination, achieving the goal of influencing the direction of public opinion.
In short, algorithmic recommendation is by no means a value-neutral technology; rather, it builds a bridge between values and users, becoming a carrier of value dissemination. "Every technical architecture, every line of code, and every interface represents a choice, implies a judgment, and carries values." The process of a user acquiring information through algorithmic recommendation is simultaneously a process of technical occupation and value output by the algorithm designer toward the user. The designer's choices and judgments will imperceptibly enhance the discourse power of their own values, transforming and reshaping the values of the algorithm user.
II. The Double-Edged Sword: The Impact of Algorithmic Recommendation on Cyber Ideological and Political Education
For cyber ideological and political education, algorithmic technology has a dual impact, providing both rare development opportunities and severe challenges. Ideological and political educators need to seize opportunities and meet challenges, accurately identifying, scientifically responding to, and actively seeking change to promote the in-depth development of cyber ideological and political education.
1. Opportunities brought by algorithmic recommendation to cyber ideological and political education
Based on precisely grasping the ideological dynamics of network users, algorithmic recommendation can achieve the precise matching and delivery of massive online information, thereby providing new development opportunities for the conduct of cyber ideological and political education.
First, by precisely delivering personalized information, human subjectivity becomes more prominent. Algorithmic recommendation is oriented by user needs, achieving personalized information delivery for different users. This effectively stimulates user excitement points and highlights the subjective existence of the individual, creating conditions for the in-depth conduct of cyber ideological and political education. Firstly, algorithms provide the fundamental prerequisite of "attraction" for conducting cyber ideological and political education. Attraction is the primary factor; without it, there is no object of cyber ideological and political education, no subject to speak of, and certainly no formation of an interactive relationship or the unfolding of interactive activities between the subject and object. In cyber ideological and political education activities, netizens possess strong independence and significant autonomy; if they lack interest in the educational content, they can choose to go "offline" at any time to end the activity. Algorithms can fully reflect the interests of netizens and effectively attract their attention, allowing ideological and political work to proceed smoothly. Secondly, algorithms provide technical support for realizing the "subjectification of the object" [5] in cyber ideological and political education. Relying on algorithmic recommendation allows for the precise grasping of and alignment with the internal needs of the object, fully mobilizing their initiative and highlighting their subjectivity. The effectiveness of cyber ideological and political education is realized precisely through this process of "the guest becoming the host." [6] Finally, algorithms help enhance the subjectivity of the subject of cyber ideological and political education. Algorithmic recommendation can help educators liberate themselves from repetitive, formulaic, and foundational tasks, giving them more time to improve themselves and creating conditions for achieving their own free and comprehensive development. On this basis, educators can use a better mental state and stronger professional skills to complete those more challenging and creative ideological and political education tasks that algorithms cannot undertake, better fulfilling the tasks of ideological guidance and value leadership.
Second, by precisely matching information with people, the acquisition of information becomes more convenient. Algorithmic recommendation enables users to quickly obtain the network information they need, laying a solid foundation for conducting cyber ideological and political education. On the one hand, algorithmic recommendation helps the subject of education better achieve a "broad vision." [7] Doing ideological and political work well requires multidisciplinary professional knowledge. Algorithmic recommendation allows the subject to quickly find various knowledge resources needed for education; reflecting on and internalizing these based on an extensive possession of knowledge resources helps improve their ability to guide cyber ideological and political education, thereby providing logical and persuasive explanations for theoretical and practical problems. On the other hand, algorithmic recommendation helps the object of education better achieve "value cultivation." Cyber ideological and political education is not hollow value preaching; the generation of any values must be built on a certain foundation of knowledge reserves. By pushing information with the user as the center, algorithmic recommendation can fully stimulate the initiative of the object to acquire information, study theory, and enlighten their thoughts, helping them better understand and grasp the correct worldview, outlook on life, values, and methodology. Supported and nourished by scientific theory, the object's value judgments will be more accurate, comprehensive, and scientific, and value leadership work will be more stable and lasting.
Third, by precisely grasping ideological dynamics, the advancement of work becomes more effective. Algorithmic recommendation can present human thoughts in the form of data through technical means, making it possible to precisely analyze and grasp ideological dynamics. In the current era, the independence, volatility, and diversity of people's ideological activities have become increasingly prominent. Realistic features such as distinct personalities, pluralistic values, and diverse pursuits have become more evident, making it increasingly difficult to grasp ideological dynamics. To achieve real effectiveness, ideological and political education must follow the principle of "using one key to open one lock," [8] providing targeted guidance and help based on a precise grasp of the object's personality characteristics, ideological patterns, and the crux of their problems. With the aid of algorithms, educators can comprehensively master behavioral data left by netizens online—such as browsing, likes, and comments—and analyze it to understand ideological dynamics to a certain extent, grasping the patterns of their thoughts and formulating personalized work plans, thereby making cyber ideological and political education more effective.
2. Challenges brought by algorithmic recommendation to cyber ideological and political education
While algorithmic recommendation technology provides new developmental opportunities for online ideological and political education, it also presents numerous problems and challenges that require profound reflection and scientific responses.
First, "technology-driven" processes cause the attenuation of the educational subject. With the aid of big data, algorithmic recommendation deeply participates in the human activity of transforming the world, altering the relationship between humans and tools to a certain extent. In online ideological and political education, algorithms can "automatically" push information to the educated, replacing the educator in performing a portion of ideological and political education activities. Does this mean algorithms have attained a subjective status equal to humans, such that a human-machine binary coexistence has become the new modality of the subject in online ideological and political education? Regarding this question, we must see, on the one hand, that algorithms do not yet possess full self-awareness; what they ultimately execute are still human instructions, reflecting the intentions of the designer—the human. In essence, the algorithm is the result of the objectification of the human as a subject. Therefore, the algorithm is not an independent subject of ideological and political education, but rather a concrete manifestation of "essential human power." On the other hand, we must also recognize that compared to traditional technical tools that completely follow human will and instructions, algorithms possess strong creativity and relative independence. Algorithms can continually update themselves based on the designer's initial intent, analyzing and grasping an individual’s personalized needs to autonomously select the information content to be pushed. If we stubbornly insist on viewing algorithms merely as traditional instrumental "objects," it will be difficult to provide a rational explanation for many new problems arising in practice. For instance, with the rapid development of bionic technology, humans may develop special emotions toward certain algorithm-based AI products (such as humanoid robots); if these products are simply categorized as instrumental "objects," it would undoubtedly face significant skepticism. Therefore, one might consider granting algorithmic tools a kind of "intelligent subject" status, allowing them to occupy a position between humans and tools.
Once algorithms intervene in online ideological and political education, they profoundly change the mode of interaction, transforming the structure of the relationship from the traditional "educational subject–educational object" model to an "educational subject–algorithm–educational object" model. The educational subject no longer faces the educational object directly but uses the algorithm as a medium to impact and influence the educational object in an extremely hidden and indirect manner. Within the framework of implementing the educational subject's intentions, the algorithm, as an "intelligent subject," plays an auxiliary educational role in the specific information-pushing process. It becomes difficult for the educational object to perceive the existence of the educational subject through the algorithm. The manifestation of the educational subject's "virtual presence" is further intensified, hindering the subject from playing its proper role and bringing new challenges to online ideological and political education. For example, virtual presence means the educational subject can mostly only exercise control at the source, while effectively managing the process and results becomes difficult. The educational subject can design an algorithmic model according to their intentions, but due to the uncontrollability of algorithmic technology, the recommended content may ultimately diverge from the subject's original expectations. Virtual presence also makes it possible for educational subjects to lower their requirements for subjective responsibility due to the anonymous nature of the online environment.
Second, "decentralization" exacerbates ideological risks. In the internet era, technical empowerment has made the structure of information dissemination increasingly flat. Everyone can become a publisher of online information, changing the traditional model of centralized production and targeted transmission. The production and release of information increasingly demonstrate a "decentralized" trend, bringing many risks to the dissemination of the mainstream ideology. First, mainstream media faces the risk of marginalization. With its advantages of efficiency, personalization, and precision, algorithmic recommendation can capture more netizens' attention, triggering a profound revolution in the landscape of information dissemination. Consequently, the space for the survival and development of mainstream media has been severely squeezed; the transmission and influence of mainstream ideology face severe challenges, and its authority and control have diminished. Second is the problem of the "vulgarization" of values. Driven by commercial interests, algorithmic platforms center content delivery and pushes on traffic growth, considering only whether it is "accurate" (in matching interests) rather than whether it is "correct" (ethically or politically). The social responsibility that platforms ought to bear is ignored. Instead of resonating with and moving in the same direction as mainstream values, the online space is filled with vulgar, low-brow, and kitsch information, which seriously erodes people's spiritual worlds, leading to a "sinking" of tastes and a lowering of moral tone. Third, domestic and foreign hostile forces use technical means to interfere in other countries' political affairs. Using algorithms, foreign forces can set agendas and guide public opinion through extremely hidden technical means to deconstruct social identity, incite social contradictions, and interfere with social stability. To such an extent, when a social event occurs, it is even difficult for us to determine whether it is a spontaneous act of public expression or the complex result of external incitement and manipulation, thus increasing the difficulty of resolution. Currently, many domestic algorithmic platforms have accumulated massive user bases; if these were controlled or interfered with by domestic or foreign hostile forces, the social harm would be unimaginable.
Third, "filter bubbles" trigger social "circle" polarization. To increase user stickiness, algorithms push information according to user interests. However, for the audience, being excessively confined to personal interest areas and only receiving content that aligns with their own inclinations causes the solidification of social circles, the narrowing of horizons, and the polarization of opinions, triggering the "filter bubble" [9] effect. As the American scholar Sunstein pointed out in Republic.com (Chinese title: Information Utopia), the public's demand for information is not all-encompassing but tends to follow interests: "a communications universe in which we hear only what we choose and only what comforts and pleases us." The singular and homogenized delivery mode of algorithmic recommendation blocks the intake of heterogeneous information. Over time, it imprisons users within a "cocoon" of their own making. Long-term exposure to homogenized information and excessive convergence of opinion not only reinforces the "filter bubble" but also causes online groups to exhibit "circle" [10] differentiation. In these "circular" spaces, members interact regarding specific types of information, while exchange with the diverse outside world is drastically reduced, leading to "intra-circle homogeneity and inter-circle heterogeneity." Long-term exposure to homogenized information makes it easy for users to form "self-confirmation bias," mistakenly believing their prejudices to be truth. Especially after their views gain approval from "allies" within the circle, this group consciousness and cognitive bias are further validated and strengthened. This leads to the rejection and exclusion of other diverse information and rational viewpoints, making judgments on specific issues and society as a whole trend toward irrationality, providing fertile soil for the emergence of extreme ideologies.
Fourth, "information fragmentation" affects the holistic construction of education. To adapt to the fast-paced lifestyle of modern society, information recommended by algorithms changes with the user's interests and emotional states. Users only need to swipe the screen to switch information rapidly, resulting in a lack of necessary coherence and correlation between the pushed information, making fragmented characteristics increasingly evident. Observing, analyzing, and grasping problems in a comprehensive, systematic, and complete manner is a fundamental requirement of ideological and political education, but the prevalence of fragmentation makes it easy for people to lose this holistic vision. Lenin once compared Marxism to an organic system "cast from a single block of steel," from which "one cannot eliminate even one basic premise, one single substantial part, without departing from objective truth, without falling into the arms of bourgeois-reactionary falsehood." Without a holistic vision, it is impossible to grasp the theoretical substance of Marxism and the basic requirements of ideological and political education in their entirety. In practical life, people become accustomed to viewing problems with isolated, scattered, and one-sided perspectives, leading to the problem where "local information increases while holistic truth grows ever more distant." Algorithmic recommendation places us in a user-led new media era. In such an era, "we will have more information than ever before, but at the same time be more easily confused; we will see the truth more easily, but the truth will also be harder to obtain." The flood of fragmented information often leaves people lacking dialectical analysis and comprehensive cognition of events, focusing instead on minor details. This generates obsessive cognitions that "take the part for the whole" or "mistake the branch for the root," dissolving the integrity and effectiveness of ideological and political education.
III. The Path of Innovation: Algorithmic Recommendation Empowering the Guidance of Mainstream Values
During the 12th collective study session of the Political Bureau of the CPC Central Committee, Xi Jinping explicitly demanded: "Explore the use of Artificial Intelligence in news collection, production, distribution, reception, and feedback; use mainstream value orientations to harness 'algorithms,' and comprehensively improve the ability to guide public opinion." [17] Ideological and political education workers should re-examine their work from the perspective of algorithms, transforming according to the matter at hand, advancing with the times, and innovating according to the momentum [11]. By strengthening institutional, technical, and guiding innovation—aiming to direct, utilize, and upgrade algorithms—we can push online ideological and political education toward new development in the era of algorithmic recommendation.
1. Strengthening institutional innovation to direct algorithms We must adhere to the principle of unifying "technical empowerment" with "technical responsibility," ensuring that algorithmic platforms assume corresponding responsibilities while obtaining technical power. First, construct a coordination and guidance mechanism for online platforms. By regularly convening coordination meetings for platform heads, we can build a long-term mechanism to strengthen macro-control and guidance over the identification and recommendation of information by intelligent algorithms. This includes increasing the push and dissemination of mainstream authoritative information and dominant value-oriented information to lead the direction and optimize the landscape of online dissemination, creating a "clean and righteous" [12] atmosphere. Second, strengthen external government supervision. Through clear institutional regulations, draw clear "red lines" for algorithmic recommendation. Use the rigidity of institutional constraints to urge platforms to establish a strong sense of responsibility. They must abide by laws and regulations, respect social ethics, fulfilling their responsibility as "gatekeepers" and assuming their due social responsibility. To this end, the state has successively issued policy documents such as the Guiding Opinions on Strengthening the Comprehensive Governance of Internet Information Service Algorithms and the Provisions on the Administration of Algorithmic Recommendation of Internet Information Services. These provide clear institutional safeguards for regulating algorithms and an important institutional basis for conducting online ideological and political education. Online platforms must be guided to strictly implement these systems. Third, strengthen the internal construction of online platforms, strictly fulfilling the duty of information auditing. To perform online ideological and political education more precisely, we can further improve the real-name authentication system for algorithm users. By supplementing missing data such as age and gender, we can achieve precise identification of different groups, making recommendations more targeted. Considering that the youth are the main force of the internet and are susceptible to the erosion of harmful information, an information auditing and classification mechanism should be established to strictly filter harmful information that minors might encounter. To this end, we should strengthen publicity and training for platform managers and employees, encouraging platforms to establish internal control systems, strengthen self-discipline, and manage risks effectively.
2. Strengthening technical innovation to utilize algorithms
It is essential to persist in the principle of unifying "technical rationality" with "value guidance," taking the dissemination of mainstream values as the fundamental logic to innovate the technical mechanisms of algorithmic recommendation. Ideological and political work is the lifeline for the high-quality development of algorithmic technology. Algorithms must take greater account of the needs of ideological and political education, transcend the interest-driven pitfall of "traffic above all else," and redefine and construct algorithmic models through the logic of "values first." Through technical means, the socialist core values should be integrated into the logic of algorithmic recommendation, allowing "mass traffic" to release the "positive energy" of values. First, while respecting users' personalized needs, the system should consciously recommend web links to external heterogeneous information. This ensures algorithms cover diverse viewpoints and disseminate many voices, helping users break free from the shackles of "information cocoons" [13]. This guides the educated to broaden their horizons and cultivate a more rational and peaceful mindset, thereby resolving the issues of "circle stratification" [14] and group polarization. Second, one must become adept at maintaining the holistic perspective of ideological and political education within an environment of fragmented information. On one hand, it is necessary to persist in using the dialectical materialist viewpoint of interconnection and development to look at problems. One must learn the analytical methods of "discarding the dross and selecting the essential, eliminating the false and retaining the true, proceeding from the one to the other, and from the surface to the core" [15]. Analyzing and organizing fragmented information is crucial, as is finding the inherent essence, interconnections, and future trends of information flows on network platforms to continuously improve systemic analytical capabilities. On the other hand, it is necessary to master the laws of information dissemination on network platforms. Algorithms should be used to integrate data resources, break down information barriers, and construct a three-dimensional mode of information dissemination that combines "points and planes," enabling ideological and political education to explicate "holistic" meaning within "fragmented" information. Third, the "re-centralization" of information dissemination must be achieved. Official mainstream media must actively "embrace" algorithmic technology, excelling at "using borrowed strength to exert force" and "borrowing a boat to head out to sea" [16]. While maintaining the correct political direction, they should plan around the central work of the Party and government, major practical issues, and important temporal milestones to firmly grasp the initiative in discourse power over public opinion. By adopting promotional methods that the people love to see and hear, actively setting agendas, and guiding public opinion, the dominant role of mainstream media can be consolidated and strengthened, filling cyberspace with positive energy, the "main melody" [17], and good voices.
- Strengthen guidance innovation and optimize algorithms. Based on guiding and utilizing algorithms well, it is necessary to further optimize and enhance them, strengthen innovation in educational guidance within the algorithmic perspective, and continuously improve the power of network discourse and value guidance. First, we must promote conceptual innovation in algorithmic delivery. Algorithm users should transform their concepts, truly learning to rely on algorithms without becoming dependent on them. While enjoying the immense convenience brought by algorithms, users must clearly see that "user-centered" algorithmic recommendation still has a significant gap from satisfying "the user's true needs." It is precisely this gap between being "user-centered" and "centered on the user's true needs" that allows algorithms to potentially move from their original positive aspect—obeying and serving humanity—toward their opposite, becoming an alien force that reifies and enslaves humanity. Yet people immersed in the comfort and convenience of algorithmic recommendation are often unaware of this harm or unable to stop. To address this, algorithm users must update their concepts and dialectically grasp the dual dimensions of the reality and transcendence of algorithmic technology. On the basis of using algorithms to possess vast amounts of information, they must engage in more independent and deep thinking, actively exploring the information, knowledge, and viewpoints they truly need. Through "deep learning," they should conduct persistent and profound study and reflection, enhancing the subjectivity of algorithm application, achieving the leap from passive to active, and promoting the comprehensive improvement of their own quality, moving steadily toward the core goal of "the free and comprehensive development of the individual." Second, we must promote innovation in the identification of algorithmic delivery. Algorithm users must identify algorithmic technology in a comprehensive and dialectical manner. Algorithms mostly push shallow, fragmented, and interest-driven information, which often brings people only "instant gratification." After obtaining this "momentary pleasure," people's inner selves are frequently occupied by a greater sense of emptiness and frustration. Therefore, ideological and political education must focus on improving users' "algorithmic literacy" to ensure they are adept at identifying algorithms. Identifying algorithms as mentioned here does not require every user to master the technical logic and parameters like an expert; rather, it aims to enhance users' dialectical thinking skills. This enables them to proficiently apply the Marxist standpoint, viewpoint, and method to comprehensively and dialectically view the positive and negative impacts of algorithms. They should be able to see the development opportunities brought by algorithmic empowerment while remaining sufficiently vigilant against, and consciously overcoming, the alienation problems that algorithms may trigger. Third, we must promote innovation in the discourse of algorithmic delivery. The deep integration of algorithmic technology and online ideological and political education brings new development opportunities while forcing corresponding changes in the discourse of online ideological and political education. Online ideological and political education must focus on innovating discourse content. It is necessary to utilize algorithmic technology to promote the networked expression of mainstream values, transforming serious mainstream discourse into vivid "network discourse," thereby expanding the effective supply of discourse content and providing a "source of living water" for the construction of online discourse power. Furthermore, online ideological and political education must promote innovation in the "mode of discourse." By utilizing the technical advantages of algorithms, it should use a localized, popularized, and accessible discourse style and adopt graphic, visual, and interactive forms of presentation to effectively enhance the affinity of online ideological and political discourse, expanding the audience in the process of "teaching through entertainment." Most importantly, online ideological and political education must focus on innovating "discourse dissemination capacity." It should use algorithmic technology to analyze, predict, and grasp the behavioral preferences of netizens, realizing the personalized development and targeted delivery of discourse content. This will effectively enhance the precision of dissemination, better satisfying the spiritual and developmental needs of netizens.
(Author Bios: Luo Yuting is the Dean of the Research Institute of Ideological and Political Education at Wuhan University, and a Professor and Doctoral Supervisor at the School of Marxism; Xiao Tianle is a Doctoral Student at the School of Marxism, Wuhan University.) Network Editor: Tong Xin Source: Ideological and Theoretical Education [18], 2023, No. 10.