Artificial Intelligence at the Core of Chinese Military Power: The New Paradigm of PLA Intelligentized Warfare

Abstract

This report provides an in-depth analysis of the People's Liberation Army's (PLA) ambitious pursuit of intelligentized warfare. China is not merely integrating Artificial Intelligence (AI) into its armed forces but fundamentally reshaping its military decision-making paradigm, moving far beyond mere automation.

The PLA is adopting a holistic approach to intelligentized warfare, leveraging dual-use civilian-military research departments integrated into large global state-owned enterprises ("megamergers") to bridge the critical gap between "strong artificial intelligence" (AI with human-like cognitive capabilities) and its complex military and defense applications.

At the heart of this endeavor is the development of a groundbreaking decision-making system engineered to autonomously think, reason, make, and precisely execute decisions, thereby extending and augmenting the cognitive and operational capabilities of the Chinese military's highest command.

This system is underpinned by a "Trinitarian Formula": "Large Model + Knowledge + Algorithm = New Quality Combat Power." The "Large Model" provides cognitive intelligence and creative reasoning, "Knowledge" offers professional military guidance and structured thought, and "Algorithms" ensure precise execution and optimization. This synergy is intended to transform military decision-making from an empirical art into a scientific method, generating novel combat capabilities. The report also highlights the broader strategic context, ethical implications, and the shift towards cognitive and deception warfare within China's vision for future conflict.

by Gabriele and Nicola Iuvinale

This report offers an extremely detailed and in-depth overview of the ambitions and developments in intelligentized warfare by the Chinese People's Liberation Army (PLA). It is evident that China is not only pursuing the integration of Artificial Intelligence (AI) into its armed forces but is also redefining the very paradigm of military decision-making, pushing far beyond simple automation.

The document describes a holistic approach to intelligentized warfare that goes beyond the mere adoption of advanced technologies. The PLA, through dual-use civilian-military research departments integrated into large global state-owned enterprises (the so-called "megamergers"), is striving to bridge the gap between "strong artificial intelligence" (AI with human-like cognitive capabilities) and its military and defense applications.

The true core of this initiative, the most revolutionary and strategic element, is the development of an unprecedented decision-making system. This system is designed to go far beyond simple support or automation; its ambition is to replicate and even surpass human cognitive capabilities. It is specifically conceived to think, engage in complex reasoning, make autonomous decisions, and, crucially, execute them with surgical precision. This level of autonomy and intellectual capability aims to provide unparalleled support to the highest echelons of the Chinese military, especially in high-pressure, rapidly evolving battle scenarios.

One such research institute, which presents itself as the world's first in national security and defense to have achieved "complete intelligence," is at the forefront of AI's military and defense application. It integrates military thinking, combat theories, research findings, business methods, professional tools, and large models. This strategy aligns with China's broader vision of "intelligentization" (智能化) of warfare, aiming for "complete military modernization" by 2035 and becoming a global superpower by 2049, the centenary of the People's Republic's founding.

The Strategic Context: Military-Civil Fusion and Beyond

The Chinese approach to intelligentized warfare is intrinsically linked to the doctrine of military-civil fusion (军民融合). This strategy is not merely an interchange of technology but a profound integration of industrial, academic, and research ecosystems, where innovations developed in the civilian sector (such as AI, big data, 5G, robotics) are systematically transferred and adapted for military purposes. This allows the PLA to accelerate the development of cutting-edge military capabilities by leveraging the enormous innovative potential of China's technological sector.

A recent example of this fusion is the use by Chinese research institutes affiliated with the PLA of open-source AI models like Meta's Llama 2 to develop militarized artificial intelligence tools, such as ChatBIT. This demonstrates how China is willing to leverage even foreign-developed technologies, adapting them for its own defense needs, including intelligence gathering and analysis, and decision support.

The Trinitarian Formula: "Large Model + Knowledge + Algorithm = New Quality Combat Power"

The analysis highlights a fundamental formula that synthesizes the new Chinese approach to intelligentized warfare: "Large Model + Knowledge + Algorithm." This triad represents an organic fusion of thinking capabilities, professional knowledge, and executive precision, with the aim of transforming military decision-making from an "empirical art" into a "scientific method."

Large Model: Cognitive Intelligence

The Large Model, which can be likened to modern Large Language Models (LLMs) and multimodal models, is the component that provides the breadth and flexibility of thought. It performs several crucial functions:

• Human-machine interaction interface: Enhances decision-making efficiency in high-pressure battle environments by allowing interaction via natural language to request information, analysis, or issue commands. The ability to grasp the essence of a problem and present clear results with visualizations is fundamental.

• Cognitive comprehension engine: Processes and integrates heterogeneous data from multiple global sources and sensors, transforming them into meaningful information for the battlefield, even when fragmented or unstructured. This includes the ability to understand the intrinsic connections between various data types.

• Knowledge fusion platform: Achieves a unified and usable representation of interdisciplinary military knowledge (tactics, strategy, logistics, intelligence), building a coherent knowledge system that improves coordination and precision in joint operations. Its ability to absorb a wide range of knowledge systems, including military science, international relations, and geopolitics, allows it to connect seemingly unrelated information to reveal opportunities or threats.

• Strategic analysis center: Utilizes vast scientific knowledge and contextual reasoning capabilities to conduct multi-level analyses of complex situations. It can identify hidden strategic patterns and weak signals, and perform "hypothetical reasoning" to simulate battlefield evolution based on different strategic choices, helping commanders understand the long-term impact of decisions.

• Decision support system: Provides recommendations based on historical cases, professional theories, and the current situation, with the ability to explain the logic of its proposals and assess risks and uncertainties. Its generalization and creative reasoning capabilities allow it to tackle unprecedented battle scenarios, transferring knowledge to new environments and proposing rapid, unconventional response strategies, moving beyond traditional military dogma.

Knowledge: The "Dao" (Way) of Military Thought

Knowledge represents the "Dao" (the Way), the professional guidance that transforms the Large Model's general intelligence into specialized military intelligence. It is not merely a collection of information but a structured and systematic mechanism for guiding thought. Its primary functions include:

• Conceptual system support: Provides a unique conceptual framework for the military sector, including classification of operational levels (strategic, operational, tactical), categorization of battlefield elements (enemy, own forces, terrain, weather), and dimensions of military capability analysis. This enables the Large Model to understand and reason within the appropriate military context.

• Operational logical relationship network: Encodes complex causal chains, influence mechanisms, and evolutionary laws relevant to military decisions, ensuring that the AI's reasoning is consistent with scientific and warfighting principles, allowing for accurate predictions and projections.

• Systematic methodological guidance: Incorporates specific military decision-making methodologies and procedures, ensuring the Large Model follows standardized thought processes and decision pathways, similar to those of an experienced professional. This ensures the scientific rigor and consistency of the decision-making process.

• Experience-derived knowledge injection: Integrates "laws of warfare," "operational principles," and "tactical secrets" distilled from countless historical cases and practical experiences. This enables the Large Model to learn from past lessons and apply accumulated wisdom, avoiding repetition of errors and identifying winning principles.

In essence, knowledge transforms AI from a "speaking AI" into a "military-savvy decision assistant," ensuring professionalism, rigor, and reliability in the decision-making process.

Algorithm: The "Art" of Precise Execution

Algorithms are the "Art" ("Shu"), the professional computing engine that translates cognitive understanding and decisions into precise actions. They provide the executive capability and precision indispensable in a military context. Their functions include:

• Professional computation and optimization capabilities: Offer high-precision calculations and optimized solutions in specific military areas, such as ballistic calculations, target identification, and resource allocation, with superior precision compared to generic models.

• Execution efficiency: They are optimized for specific battlefield tasks, significantly improving computational efficiency where time is a critical factor.

• Determinism guarantee: Their results are verifiable, predictable, and highly reliable, as they are based on clear mathematical models and logical rules, which is crucial for high-risk military operations where reliability is paramount.

• Low resource consumption: Designed to be lightweight, suitable for deployment in resource-constrained environments like forward combat units and unmanned platforms, enabling real-time intelligent processing directly on edge devices.

• Precision control: Provide detailed control and fine-tuning capabilities for modern weapon systems and unmanned platforms, ensuring precise execution of operations and reducing collateral damage.

Algorithms serve as the "mathematical blade" of intelligentized warfare, bridging the gap between "thinking" and "doing," and ensuring that the Large Model's insights translate into precise tactical actions.

Synergy Mechanisms and Cooperative Functionality

The system's effectiveness lies not only in the individual capabilities of each component but in their synergy and division of labor:

• "Thought-Calculation" Separation: The Large Model handles cognitive comprehension and strategic thinking (uncertain and ambiguous tasks), while Algorithms manage precise calculations and tactical execution (deterministic problems). This separation ensures the system can address both the complexity and uncertainty of the battlefield while maintaining precision.

• Complete Decision Chain: A closed loop is created, from situation perception to action execution. The Large Model comprehends information, Knowledge guides analysis and preliminary plan formulation, Algorithms optimize calculations, and finally, results are integrated for the final decision. This process ensures a hierarchical decomposition from macro to micro, without interruptions.

• Complementary Reinforcement Effect: The Large Model's generalization compensates for the domain limitations of Algorithms, while the Algorithms' precision mitigates the Large Model's probabilistic flaws. The Knowledge framework enhances the Large Model's professional depth and provides Algorithms with correct application contexts. This creates a synergistic effect where "1+1>2."

• Dynamic Adaptation to Complexity: The collaboration between components is not static but dynamically adapts to task complexity. For simple tasks, Algorithms can act directly; for complex problems, the Large Model is deeply engaged; in highly specialized areas, Knowledge plays a more prominent role. This ensures efficiency in resource utilization.

The Knowledge Guidance Mechanism

Knowledge guidance is a central mechanism that transforms static knowledge into dynamic thought guidance.

• Knowledge Modeling: Military concepts, their relationships, and key decision-making procedures are transformed into structured thought models. This transforms knowledge from a consultable resource into an active framework for the Large Model's reasoning.

• Professionalism Enhancement: Through military terminology dictionaries and professional conceptual systems, the system is guided to use domain-specific military language and logic, ensuring analyses are conducted with a professional military mindset.

• Rigor Guarantee: The introduction of reasoning constraints, verification checkpoints, and self-consistency mechanisms ensures that generated operational plans comply with military principles and operational constraints. This makes decisions "error-proof" within established limits.

• Methodological Guidance: The knowledge framework incorporates systematic military analysis methods and decision procedures, such as objective decomposition, element evaluation, and solution comparison. This guides the Large Model to follow a scientific and explainable thought path, avoiding intuitive judgments or hasty reasoning.

Generation of New Quality Combat Power

The synergy between the Large Model, Knowledge, and Algorithm gives rise to a series of New Quality Combat Power, which goes beyond the quantitative improvement of existing capabilities:

• Knowledge-Driven Cognitive Decision Advantage: Intelligentized warfare shifts competition from mere information acquisition and transmission to deep understanding and application of knowledge. The system not only "sees" phenomena but "understands" their essence and mechanisms, providing commanders with profound battlefield insights and the ability to seize critical windows of opportunity. This represents a fundamental shift in the ability to maintain the initiative.

• Professionally Guided Decision Resilience: In the face of uncertainty and rapid changes on the battlefield, the system maintains a professional perspective and systematic thinking, providing reliable recommendations based on military laws and methodologies. This decision resilience allows the command system to remain stable even under conditions of high stress and unpredictability.

• "General + Specialized" Intelligent Integration: The system combines the general adaptability of the Large Model (for unforeseen scenarios and new types of operations) with the specialized precision of Algorithms. This enables the system to be both flexible and precise, overcoming the dilemma of traditional systems that are "specialized but rigid" or "generic but superficial." This is particularly relevant for addressing new threats like cognitive warfare or digital deception, where China aims to sabotage the opponent's information ecosystem.

• Seamless Integration of All-Domain Perception and Precise Decisions: The system achieves a seamless conversion from strategic to tactical levels, integrating macroscopic battlefield understanding with precise tactical action planning. This integration eliminates information and decision gaps between strategy and tactics, ensuring more consistent execution of operations and closer adherence of tactical action to strategic intent. This aligns with the Chinese concept of Integrated, Digitalized, and Intelligent Command, as demonstrated by the construction of cutting-edge military command centers.

• Balance Between Innovation and Norms: The system ensures that decisions conform to military laws and professional standards through the knowledge framework, while leveraging the Large Model's creative thinking to explore innovative solutions that go beyond conventional thought. This "innovation within norms" provides new strategies based on solid military scientific foundations, allowing the PLA to adapt quickly to new threats and propose creative countermeasures.

• New Military Command Paradigm Based on Deep Knowledge-Guided Human-Machine Collaboration: Knowledge serves as a common language between human experts and AI, enabling deep collaboration where commanders are not just users but members of a complementary team. AI understands implicit human intentions and reasoning logics, while commanders focus on value judgments and creative thinking, significantly enhancing command effectiveness and decision-making quality. This shifts command from an experience-based model to a more scientific and intelligence-assisted one.

Considerations and Future Implications

China's massive investment in military robotics and autonomous systems, such as the "mechanical yak" and other surveillance and combat robots, highlights the urgency with which the PLA is striving for technological superiority. Experts suggest that China could be among the first nations to employ fully autonomous weapon systems in active operations, which would represent a paradigm shift in the international security architecture.

However, the integration of AI systems into military decision-making also raises significant ethical and control challenges. Research has shown that large language models, when used in simulated war games, can tend towards escalation, even to the use of nuclear weapons. This highlights the critical need for effective human oversight and robust governance, issues that China itself has acknowledged within the "Global AI Governance Initiative."

The doctrine of intelligentized warfare is therefore not limited to the development of new weapons but extends to cognitive warfare and "deception warfare," where the goal is to pollute the opponent's information ecosystem with false or distorted data, rendering their decision-making systems blind.

Ultimately, China's intelligentized warfare strategy is an ambitious project aimed at redefining how wars are fought and won, placing AI at the core of the decision-making process and operational execution. Its implications are vast and warrant continuous attention from the international community.