Zhu Qichao: AI Empowerment of National Security: Current Status, Trends, and Countermeasures
Technological empowerment is the key support for advancing the modernization of the national security system and capabilities. The Third Plenary Session of the 20th CPC Central Committee proposed: "Build a coordinated and efficient national security protection system and promote technological empowerment for national security." During the 20th collective study session of the Political Bureau of the 20th CPC Central Committee, General Secretary Xi Jinping emphasized: "We must face the gaps squarely, redouble our efforts, comprehensively promote AI technological innovation, industrial development, and empowering applications, improve AI oversight systems and mechanisms, and firmly grasp the initiative in AI development and governance." As a key technology possessing strategic value, general-purpose attributes, and dual-use characteristics, AI is accelerating its penetration into various fields through the transformation of information processing paradigms, innovation in data analysis methods, optimization of assisted decision-making models, and expansion of security governance boundaries. It is profoundly reshaping the national security landscape and even the strategic posture of international gamesmanship. To comprehensively implement the Holistic Approach to National Security [1] and advance the modernization of the national security system and capabilities, we must attach great importance to the "AI Chapter" of technological empowerment for national security. We must accelerate the application of frontier and disruptive technologies represented by AI, update work concepts, design interconnected scenarios, promote the implementation of results, and clarify governance norms. This will inject new kinetic energy into every field, system, and element of the national security system, building a resilient line of defense for national security and dynamic development.
The Basic Logic of AI Empowering National Security
The world today is in a critical stage of transitioning from industrial civilization to digital civilization. While the accelerated development of AI brings new risks and challenges to national security, it also brings unprecedented opportunities for empowered development. Although the security narrative of great power competition under the traditional geopolitical perspective continues, emerging security factors are rising strongly, and a new security logic is quietly forming. Existing research exploring the relationship between AI and national security primarily follows two theoretical paths: first, "technological instrumentalism," which views AI as a tool to improve the efficiency of security governance; and second, "risk-originalism," which emphasizes that AI itself constitutes a new type of security threat. Neither path fully explains the basic logic of AI empowering national security. This article argues that we should establish a "holistic security view," analyzing how AI drives the transformation of national security governance paradigms and the profound evolution of the international strategic landscape from three dimensions: technological empowerment, institutional gamesmanship, and the reshaping of power.
Technological empowerment: the robust demand for intelligent governance systems. China's 14th Five-Year Plan points out the need to "strengthen the construction of a digital society and digital government, and improve the level of digitalization and intelligence in public services and social governance." Through the synergistic effect of algorithms, computing power, and data, AI has broken through the temporal-spatial limits and cognitive boundaries of traditional national security governance. It promotes the transformation of the security system from passive response to proactive prediction, forming an intelligent governance system based on the "data-algorithm-computing power" synergy. This new governance paradigm builds a domain-wide situational awareness network through multi-modal sensing technology, relies on machine learning for pattern recognition and risk assessment of complex threats, and uses autonomous decision-making systems to complete the dynamic optimization of response strategies. Its characteristics are manifested as: governance timeliness achieving a leap from lagging disposal to real-time early warning; governance dimensions expanding from traditional physical space to digital and cognitive spaces; and the governance mode shifting from human-led to the highly efficient operation of human-machine coordination characterized by informatization and objectification. This technology-driven governance transformation not only renovates the technical architecture of the national security system but also gives rise to new governance concepts such as algorithmic governance, resilient governance, and cross-domain governance, marking the entry of national security governance into a new developmental stage with data intelligence as the core driving force.
Institutional gamesmanship: the dynamic balance between algorithmic autonomy and political governance. The deep application of AI is restructuring the power structure of national security governance. Its core contradiction lies in how to handle the dynamic balance between the expansion of algorithmic autonomy and the control of the political system. Traditional governance modes face a triple decision-making dilemma: first, the physiological limitations of human cognitive ability make it difficult for decision-makers to effectively process massive information flows, forming an "information overload" dilemma; second, the filtering effect in bureaucratic information transmission causes the fragmentation and distortion of decision-making bases—"information silos" [2] have not only failed to be alleviated in the era of informatization and intelligence but have, in some sense, been continuously reinforced; third, the experiential dependence and groupthink tendencies of decision-making subjects are prone to producing systemic cognitive biases. Together, these structural defects constitute the "bounded rationality trap" of traditional decision-making, leaving the governance process continuously exposed to risks of information asymmetry, errors in judgment, and delayed responses. On the one hand, technological empowerment can compensate for the blind spots in traditional decision-making processes, breaking the "information silos" or "information cocoons" between decision-making subjects, making algorithms a key tool for assisted decision-making. On the other hand, the political system must protect technological autonomy, national sovereignty, and security through institutional innovation, establishing adaptive mechanisms such as algorithmic review and dynamic oversight. This process of institutional adjustment is, in essence, a dynamic game between technical rationality and political governance rationality, ultimately forming a new security governance architecture characterized by the coordination of humans, machines, and the environment.
Power reshaping: a key variable in the evolution of the international power structure. At the level of international relations, AI has become a key variable in reshaping national security capabilities and the international power structure. The principle of sovereign equality is a basic principle of modern international law, emphasizing that regardless of differences in size and development, all sovereign entities should enjoy equal rights to international participation. However, the realistic operational logic of international politics shows that great powers, by virtue of their comprehensive national strength, naturally occupy a more dominant discourse power and the ability to shape rules in international affairs. Furthermore, when structural changes occur in the balance of power between great powers, the international political landscape also changes. As a key variable reshaping the international strength landscape, AI can significantly amplify the power gap between state actors, creating asymmetric competitive advantages for technological leaders. Tech-leading countries, especially hegemonic powers, often use AI advantages to strengthen their proactive posture in strategic gamesmanship and strategic deterrence, while technological dependence may exacerbate the security vulnerability of technologically backward countries. This technology-driven power reconstruction mechanism enables countries that master AI advantages to achieve "overtaking on the curve" [3] in international competition, thereby changing the existing international power distribution pattern. This makes AI not only an empowering tool for national security but also an intermediary variable for the redistribution of international power, promoting the formation of a new global security order under the background of the politicization of technology.
The Current Status of AI Empowering National Security Development
Currently, advanced technologies represented by AI are empowering and penetrating various fields of national security with unprecedented breadth and depth. This both injects new kinetic energy and introduces new governance means to various fields of national security, while also causing the internal connotations, boundaries, and outward development of national security to face new requirements of the times. From the perspective of technological empowerment, AI is playing an important role in enhancing intelligence perception capabilities, assisting decision-making, and updating scenario-based means. From a practical perspective, major countries are stepping up the development and application of new AI system tools to provide new support for national security work under the new situation.
Enhancing intelligence perception capabilities. Currently, national security intelligence work is facing a triple challenge: explosive data growth, increasingly complex information system structures, and more diverse application scenarios. By developing multi-source heterogeneous data processing capabilities, modern AI systems can achieve real-time fusion analysis of diverse information flows such as social media metadata, geospatial imagery, and open-source intelligence. Such technology can significantly improve the accuracy and timeliness of threat identification, increasing analysis efficiency to several times that of traditional methods. In particular, the breakthrough progress of generative AI, through the establishment of multi-modal correlation models, has not only restructured the production method of intelligence products but also spurred new analysis paradigms such as dynamic threat deduction. Intelligence judgment systems based on advanced algorithmic models can significantly shorten the strategic situation assessment cycle. This leap in capability is reshaping the time window and response mechanisms for national security threat assessment. For example, the "Gotham" system developed by the American AI company Palantir to address the cross-departmental "data silo" problem in the US intelligence community after the "9/11" incident is a national security-grade data analysis system. Its core technical architecture is supported by a distributed data fusion engine, supporting real-time access to multi-modal data streams generated by sensor networks, satellite remote sensing systems, and various monitoring terminals. Through self-developed semantic parsing algorithms and spatio-temporal correlation models, the system achieves a three-dimensional visual integration of unstructured text, geospatial coordinates, and real-time imagery. Its decision support system can automatically generate threat assessment maps and provide dynamic operational plan deductions for commanders. The intelligence processing efficiency of this system in counter-terrorism operations has been significantly improved compared to traditional systems, and it has now become one of the most influential intelligence analysis systems in the global security field.
Optimizing assisted decision-making. At the level of assisted decision-making, AI's empowerment of national security is mainly reflected in three dimensions: cognitive enhancement, decision optimization, and risk control. At the cognitive level, through multi-modal data fusion and knowledge graph construction, AI systems can break through human cognitive limitations and establish correlation networks of complex security threats. At the decision-making level, dynamic deduction models based on machine learning can generate joint action plans, significantly improving the scientific and forward-looking nature of decision-making. At the risk control level, AI's real-time monitoring and prediction functions help the national security system build a more agile threat response mechanism. For instance, the US Defense Advanced Research Projects Agency (DARPA) has in recent years launched an AI project codenamed "Deep Green," which attempts to propose optimal solutions for commanders to choose from by sorting through past data and integrating present data. This project has now evolved into the US Pentagon's "Intelligent Assistant to the Chief of Staff." In recent years, during the "Project Convergence" series of exercises, the US military has focused on four AI systems for tactical intelligence target access nodes—the "Rainmaker" system, "Prometheus" system, "Firestorm" system, and "Shot" system. These can help the US military quickly form an assisted decision-making kill chain from multi-sensor reconnaissance to optimized precision weapon strikes, shortening the time from target discovery to hit from several hours to less than one minute.
Enriching scenario-based means. In fields related to national security, AI has brought more abundant and efficient scenario construction and security response means. For example, in terms of national land security monitoring and early warning, intelligent monitoring systems based on computer vision and sensor fusion have broken through geospatial limitations, achieving real-time perception and automatic identification of abnormal border activities. Through machine learning algorithms, multi-source heterogeneous data can be correlated and analyzed to build dynamic threat assessment models, significantly improving the predictive judgment capability for security threats. Intelligent decision support systems can optimize resource allocation paths, forming a closed-loop security disposal mechanism. In the field of military security, technological progress and wartime needs have prompted the application scenarios of unmanned combat platforms to become increasingly diversified, with their functions developing from single use to composite and combined applications. At the same time, related technological breakthroughs will drive unmanned combat platforms to evolve from a "remote control" primary mode to a "scenario-adaptive" mode. For example, the "Autonomous Warrior" exercise jointly held by the US, UK, and Australia deployed more than 30 types of unmanned systems, including drones, ground unmanned vehicles, and underwater robots, covering various combat scenarios such as reconnaissance, strike, and logistics. The US Marine Corps recently provided the Special Operations Command with a new type of four-legged unmanned ground robot, the Vision 60. Equipped with a rifle and an AI-driven target detection system, this robot can identify enemy targets and replace soldiers in military tasks. South Korea plans to form 90 AI combat pilot units by 2028, focusing on developing unmanned platforms adapted to urban street fighting and mountain warfare. Furthermore, in the Ukraine crisis, the American AI company Palantir provided the Ukrainian military with powerful software systems, using AI to achieve precise image recognition. By performing correlation analysis on images from satellites, drones, and social media channels, it provided important technical support for the Ukrainian military in target identification, targeting strikes, wartime early warning, and assisted decision-making. The CEO of Palantir stated: "The power of warfare systems relying on advanced algorithms is equivalent to using tactical nuclear weapons against conventional weapons."
Protecting critical infrastructure. In the field of critical infrastructure protection, the application of AI technology shows a significant...
The characteristic of "securitization" has improved the risk-response paradigm of traditional infrastructure through algorithmic governance. Its technical empowerment is mainly manifested in three dimensions: for threat perception, anomaly detection systems based on deep learning have achieved multi-modal, real-time monitoring of cyberattacks and physical destruction; for risk assessment, vulnerability analysis models of critical nodes constructed via Graph Neural Networks (GNNs) can dynamically predict the transmission paths of systemic risks; and for emergency response, autonomous decision-making systems have formed a closed-loop management mechanism for self-healing and resource scheduling. This transformation of technical governance not only enhances the resilience level of infrastructure but also drives the logic of security to realize a paradigm shift from passive defense to proactive prediction. Simultaneously, it highlights the paradox between technical dependency and systemic vulnerability—the enhancement of algorithmic autonomy reinforces protective capabilities, yet may also trigger new systemic risks due to the "black box" nature of the technology, reflecting the complex dual nature of security governance in the AI era. In August 2024, the U.S. Defense Advanced Research Projects Agency (DARPA) held the semifinals of the AI Cyber Challenge, seeking to use generative AI technology to develop fully automated cyber reasoning systems to rapidly identify and repair vulnerabilities in software, systems, and networks of critical infrastructure. The report Securing Critical Infrastructure in the Age of AI, released by the Center for Security and Emerging Technology (CSET) in late 2024, emphasized that despite certain security risks, AI still possesses significant advantages in the protection of critical infrastructure.
Trends in the Development of AI-Empowered National Security
A new wave of AI development and application is sweeping the globe and remains in the ascendant, bringing precious development opportunities to all countries. However, certain major powers, out of self-interest, obsessively treat AI as a tool to maintain their hegemony, attempting to set barriers for other countries in AI development and application to create "small yards with high fences" [4] or an "intelligence divide." In this context, understanding the empowering interaction between AI and national security requires a systematic analysis from four dimensions: technological evolution, governance transformation, conceptual innovation, and international competition. Looking toward future development, AI will not only change the means of maintaining national security but will also redefine the connotation and extension of "national security" itself. As algorithms begin to participate in national security decision-making, data increasingly becomes a strategic resource, and computing power constitutes the foundation of national power, we will witness a New Era of deep mutual construction between technical politicization and the intelligentization of governance.
Dual iteration of technological evolution. From the dimension of technological evolution, AI systems are undergoing a shift from discrete tool applications to an integrated intelligent ecosystem. This evolution process is mainly reflected in three levels: At the algorithmic fusion level, cross-innovation in frontier algorithms such as deep learning, reinforcement learning, and federated learning is breaking traditional technical boundaries. Deep learning achieves the automatic extraction of complex data features through multi-layered neural networks; reinforcement learning grants systems the capacity for autonomous decision optimization in dynamic environments; and federated learning resolves privacy protection issues in collaborative computation involving multi-source heterogeneous data. This algorithmic synergy allows AI systems to construct a complete cognitive closed loop of "data collection - feature extraction - pattern recognition - decision output." At the system integration level, new-generation AI technologies are empowering the operational modes of national security systems. The intelligence collection phase achieves all-domain scanning through multi-modal sensory networks; the threat assessment phase utilizes knowledge graphs and causal inference technology to establish cross-domain correlation analysis; and decision support systems, based on deep reinforcement learning frameworks, form dynamic optimization strategies in scenarios such as military command and counter-terrorism early warning. This multi-domain integration and cross-domain fusion of situational information provides national security systems, for the first time, with near-real-time global situational awareness and second-level autonomous response capabilities. However, this trend toward technical integration also generates new governance dilemmas. The exponential growth of algorithmic complexity makes the "decision black box" problem increasingly prominent; the non-explainability of deep neural networks may mask systemic biases; the collaborative evolution of multi-agent systems may produce unpredictable emergent behaviors; and the risk of gradient leakage in federated learning poses a potential threat to data sovereignty. These challenges essentially reflect the deep contradiction between technical advancement and systemic reliability, which urgently needs to be balanced through emerging technical means such as Explainable AI (XAI) and formal verification. Future technological development must balance performance breakthroughs with security and controllability to ensure the sustainable development of an integrated intelligent ecosystem.
Accelerated transformation of governance models. Regarding the transformation of governance models, the application of AI is driving national security decision-making mechanisms to evolve from traditional "anthropocentrism" toward "human-machine collaborative symbiosis." This transition is primarily reflected in three levels: In terms of decision-making processes, traditional national security decisions relied on "expert experience + limited data analysis," but the intervention of AI has evolved the decision system into a three-tier architecture of "data-driven + algorithmic optimization + human supervision." Intelligent systems have evolved from auxiliary tools into "decision-making subjects" in certain scenarios, directly participating in operational decisions for key links, such as automated threat blocking in cyber offense and defense or dynamic troop deployment in military command. In terms of cognitive modes, AI systems can break through the physiological and logical limitations of human cognition through multi-modal correlation analysis and complex systems modeling. Machine learning can identify patterns difficult for humans to perceive, effectively avoiding "confirmation bias" and "groupthink" in human decision-making. In terms of response mechanisms, autonomous systems possess real-time situational awareness, dynamic strategy adjustment, and tiered response execution capabilities, giving the national security system the distinct characteristic of continuous evolution. This governance transformation is redefining the organizational forms and operational logic of national security.
Mutation and transition of security concepts. From the dimension of security concepts, AI is driving a mutation in national security governance philosophy, which is essentially the product of the dual mutual construction of technical politicization and the intelligentization of governance. This shift is mainly reflected in three key dimensions: First, at the level of security subjects, algorithms have transcended their status as tools to evolve into new types of weapon systems. In 2017, the U.S. Department of Defense proposed the concept of "algorithmic warfare," marking a transition in the form of warfare toward cognitive domain confrontation centered on intelligent algorithms; algorithmic advantage has become a new strategic high ground following air and sea superiority. Second, in terms of geopolitics, the struggle for dominance over digital space has birthed new fields such as "digital geopolitics." To maintain its absolute security in digital geopolitics, the United States issued the CHIPS and Science Act and established the "Chip 4" alliance, leading to increasingly fierce digital geopolitical strategic jockeying. Finally, in terms of governance logic, the deep application of AI decision-making systems has brought about the "algorithmic governance paradox": while it enhances the precision of security decision-making, the technical black-box nature puts traditional political accountability mechanisms at risk of failure.
Intensification of international competition. While AI technology is undergoing rapid iterative updates, the global governance system faces a clear dilemma of "regulatory lag," which will affect the overall effectiveness of national security governance. That is, it is increasingly difficult for a single country's national security to remain isolated; more attention must be paid to the benign interaction and coordinated governance of domestic and international security in emerging fields like AI. Global AI security governance is a frontier issue and a thorny problem in global technical governance; AI technical standards and governance rules have become a new high ground for strategic jockeying among major powers. The struggle among major countries over AI technical R&D, data governance, and algorithmic standards is, in essence, a competition over dominance of the future international order. This competitive state both drives the rapid development of AI and exacerbates the risk of fragmentation in international security governance. Future development trends will depend on the interaction of two key variables: first, the degree of synergy between technological innovation and governance innovation, requiring an agile governance framework to adapt to rapid technological evolution; and second, the progress of the international community in building a consensus on AI security governance, including coordination of technical standards, cooperation in risk management, and the formulation of ethical norms. Against the current backdrop of intensified strategic competition between major powers, global digital governance beyond the level of regional cooperation is almost in a state of total stagnation. Only by achieving a benign interaction between technology and governance can the process of AI-empowering national security maintain innovative vitality while operating on a track of safe and controllable development. This process concerns not only the security interests of individual states but will also profoundly influence the evolutionary direction and constructive effectiveness of the global security governance system.
The Path for AI-Empowered National Security Development
General Secretary Xi Jinping noted: "Accelerating the development of a new generation of artificial intelligence is an important strategic handle for us to win the initiative in global technological competition, and an important strategic resource for promoting the leapfrog development of our country's technology, the optimization and upgrading of industries, and the overall leap in productive forces." On one hand, in the development and application of AI, China faces rare opportunities of the times. AI is not only a key breakthrough for China to achieve leapfrog development in important technological fields but also provides strategic support for maintaining national security in the New Era. On the other hand, while the deep application of AI technology enhances national security capabilities, it may also bring multi-dimensional new security risks. These risks stem from the inherent uncertainty of the technology itself and the lag in existing governance mechanisms for responding to emerging technical risks. Therefore, while actively promoting the innovative development of AI technology, China must focus on constructing a corresponding risk prevention and control system to achieve a dynamic balance between technological innovation and the provision of security.
Cultivate an independent and controllable AI security technology ecosystem. At the heart of the technical path for a system of AI security empowerment with Chinese characteristics lies the establishment of an independent and controllable technology ecosystem, ensuring the autonomy of key algorithms, computing power, and data. It is recommended to focus on breakthroughs at three levels: at the foundational support level, strive to overcome "bottleneck" [5] technologies such as chip manufacturing and algorithmic frameworks to achieve autonomy and controllability of core technologies; at the algorithmic R&D level, focus on developing intelligent algorithms with explainability and anti-interference capabilities to ensure the safety and reliability of intelligent systems in auxiliary decision-making; and at the application innovation level, promote the deep integration of AI with new-generation information technology to build a foundation for an intelligent security protection system. Simultaneously, attention should be paid to military-civilian synergy in technical R&D, forming a closed-loop innovation system of "requirement pulling - technical tackling - industrial landing," so that AI security technology can both effectively safeguard national security and drive the high-quality development of industries related to new quality productive forces.
Continuously enrich the vivid scenarios of national security empowerment. With the continuous development of science and technology and the expanding map of national interests, the connotation and extension of national security are constantly breaking through original boundaries and frameworks. AI technology demonstrates broad application prospects in the field of national security, stimulating and driving the continuous evolution and expansion of the supply-and-demand landscape for national security. Generally speaking, the application of AI in the national security field is characterized by "deep technical penetration, all-domain scenario coverage, and intensified innovation competition." In the process of promoting the modernization of the national security system and capabilities, China must, on one hand, focus on constructing security scenarios to pull the demand side for AI-empowered applications; on the other hand, China must focus on using advanced technologies like AI to empower and transform original national security-related infrastructure, governance means, and standards and norms, providing new kinetic energy for the upgrading of national security system capabilities.
Establish an agile governance institutional framework for AI security. Constructing a systematic and complete institutional system for technological innovation is the strategic cornerstone for fortifying AI technical security and national security. AI security governance requires the construction of a dynamically balanced institutional framework, achieving the coordinated advancement of security and development through systematic institutional innovation. Regarding risk management, a graded supervision system based on application scenarios should be implemented, establishing a "negative list" management mechanism for core areas involving national security. Regarding innovation incentives, it is suggested to promote a "secure and controllable" technical testing mechanism to reserve institutional space for frontier technological exploration. Regarding legal protections, a full-chain institutional system covering algorithmic auditing, data governance, and ethical review should be actively constructed. By establishing a cross-domain collaborative governance architecture and resilient, adaptive policy tools, an institutional environment can be formed that both effectively prevents systemic risks and continuously stimulates innovative vitality, ultimately achieving a fundamental transition in AI security governance from passive response to proactive shaping.
Build a multi-track AI security governance community. Under the background of globalization, China needs to actively participate in shaping international rules for AI security governance and promote the formation of a more just and reasonable new international governance order. First, China should uphold...
Under the global governance outlook of "extensive consultation, joint contribution, and shared benefits" [6], China should rely on multilateral consultation mechanisms within the United Nations framework to contribute "Chinese solutions" on major issues such as the control of artificial intelligence (AI) militarization and the definition of data sovereignty. Secondly, it is necessary to establish a tiered and classified trust system for technical cooperation, constructing "buffer zones" for risk management through bilateral dialogue mechanisms such as China-US and China-EU relations, while simultaneously deepening technical security cooperation based on multilateral platforms like the BRICS countries and the Shanghai Cooperation Organization (SCO). Thirdly, China should innovate international cooperation paths for governance models by integrating the construction of the "Digital Silk Road" [7] with AI governance, promoting the governance model of "attaching equal importance to development and security" [8] which possesses Chinese characteristics. Through this multi-track approach, China can safeguard its core interests in the formulation of international rules for AI while promoting the construction of a new "non-zero-sum game" pattern for global AI security governance, achieving a dynamic balance between technological competition and international stability.
Conclusion
AI is driving the transformation of the national security system from an industrial civilization paradigm to a digital civilization paradigm through a triple mechanism of "technological empowerment, institutional adaptation, and power reconstruction." This transformation is manifested not only in an efficiency revolution in fields such as intelligence perception and strategic decision-making but also in the evolution of the underlying logic of security governance. When algorithms begin to participate in warfighting decisions, data becomes a strategic asset, and computing power constitutes the foundation of power, traditional security boundaries and governance rules face systemic reshaping.
For our country, this transformation represents both a strategic window for achieving technological "overtaking on a curve" [9] and a governance test in balancing innovative development with risk prevention and control. In the future, China needs to build a solid security cornerstone with an "independent and controllable technological ecosystem," enhance governance effectiveness with a "dynamically balanced institutional framework," and shape a consensus on rules through "multi-track international cooperation," thereby carving out a path for national security development that combines Chinese characteristics with a global perspective. This is not only an inevitable choice for safeguarding core national interests but also an important practice for building a community with a shared future for humanity. As the development of frontier and disruptive technologies increasingly becomes a global public issue, only by maintaining organic unity between security and development, innovation and governance, and autonomy and openness can we promote AI to truly become a transformative force for enhancing human well-being.
(The author is the Director of the National Defense Science and Technology Strategy Research Think Tank at the National University of Defense Technology)
Source: Frontiers (Academic edition of People’s Forum), Issue 9, 2025. Web Editor: Ma Jingren.