人工智慧時代中國軍事作戰特徵簡析

現代英語：

AI is propelling warfare into an era of intelligent warfare. The winning mechanism of traditional warfare is “the big eat the small,” a contest of troop size and firepower density; the winning mechanism of information warfare is “the fast eat the slow,” pursuing a strike effect of detection and destruction; while the winning mechanism of AI warfare is “the sophisticated eat the crude”—whoever can achieve more refined perception, more accurate judgment, and more precise strikes through algorithms will gain the initiative on the battlefield. In this context, clarifying the characteristics of warfare in the AI ​​era can help us better grasp and utilize this winning mechanism, thereby gaining the upper hand and taking the initiative on the future battlefield.

Feature 1: Comprehensive Perception – From “A Fog of Darkness” to a “Transparent Battlefield”

AI is reshaping battlefield perception, achieving unprecedented battlefield transparency. Its core lies in breaking down barriers between military branches and operational domains, effectively integrating heterogeneous data streams from land, sea, air, space, cyber, and electronic domains. Through real-time fusion and processing, a unified, dynamic, and panoramic battlefield situation map is constructed. Ideally, commanders will no longer see fragmented local information, but a complete presentation of the entire battlefield situation, providing comprehensive information support for their decision-making. It is worth noting that this transparency highly depends on the purity and timeliness of the data. The party that possesses key, accurate, and real-time updated data holds the key to “see through” the battlefield; while the party that can effectively implement data pollution, inject false information, or conduct electromagnetic spectrum deception can plunge the adversary into a deeper “fog of war,” causing confusion or even blindness in their situational awareness. Therefore, we must clearly recognize that battlefield transparency brought about by AI is not equivalent to “omniscience”: while artificial intelligence can provide near-complete “battlefield pixels,” the authenticity, timeliness, and even the correlation of the information behind each “pixel” still rely on robust information verification and analysis mechanisms to ensure its accuracy.

To achieve this, a deep human-machine collaborative information analysis mechanism needs to be established. With the support of AI technology, “human + machine” becomes a new combat formation. AI acts as a high-speed, tireless “super advisor,” while humans are the ultimate decision-makers who gain a comprehensive understanding and grasp the overall direction. In this process, AI leverages its advantages of speed, accuracy, and the ability to operate at high intensity for extended periods, demonstrating its prowess in the initial screening of massive amounts of data, the automatic identification of key features, and the rapid early warning of potential threats. Humans, on the other hand, need to exercise their initiative, recognizing the “ethics of war” that AI might overlook, upholding the “political stance” that AI might ignore, and clarifying the “value judgments” that AI might overlook, ensuring that all decisions are human-led and always serve the overall strategic interests.

Feature Two: Rapid Closure of the Kill Chain – From “Hours” to “Milliseconds”

The deep integration of AI technology is driving a leapfrog upgrade in the speed of the kill chain closure in modern warfare: from the “hours” of the mechanized era and the “minutes” of information warfare to the “milliseconds” of the intelligent era. This change stems from the deep reconstruction of combat processes by algorithms, transforming the traditional linear kill chain into a dynamic response network. This enables AI to demonstrate strong battlefield autonomy, autonomously identifying high-value targets, autonomously planning optimal response plans, and autonomously allocating resources to the most suitable combat units for execution. This brings the “detect and destroy” combat objective closer to reality: once a target is captured by sensors, its coordinates, attributes, threat level determination, optimal strike plan generation, weapon platform allocation, and firepower release can be completed simultaneously in a short time, achieving an instantaneous closed loop from perception to action.

From this perspective, algorithms are weapons—code may be more lethal than missiles. Modern warfare has evolved from a contest of performance between single weapon platforms to an intelligent confrontation of entire combat systems. A smart kill chain system capable of millisecond-level closed-loop systems is powerful not only in the number of sophisticated weapons it can deploy, but also in its ability to autonomously complete the “decision-action” cycle. In this process, the autonomous decision-making logic built by code, the cross-domain collaborative network driven by data flow, and the real-time dynamic response supported by computing power constitute the core elements of generating new combat capabilities. At this point, soft-kill attacks on algorithmic systems can have strategic effects, such as disrupting enemy decision-making models, damaging enemy cloud computing nodes, and severing enemy battlefield data chains. The impact of such attacks may be even greater than destroying an enemy’s heavy arsenal with missiles. Therefore, in planning future warfare, algorithmic control must be placed on an equally important strategic position as information control and other emerging forms of control. Efforts must be made to build an intelligent combat system with millisecond-level autonomous decision-making capabilities and strong anti-interference resilience to ensure that the rhythm of combat is firmly controlled in complex adversarial environments.

Feature Three: The Boundary Between Reality and Virtuality Disappears—From the “Real Battlefield” to the “Global Battlefield”

In the era of information- and intelligent warfare, the clear-cut physical and virtual boundaries of the traditional battlefield are rapidly dissolving, gradually evolving into a “full-domain battlefield” where the physical, information, and social domains are highly integrated and interact with each other. The firepower clashes in physical space and the bit attacks and defenses in digital space are no longer two parallel lines, but rather a unified whole that is intertwined, mutually shaping, and deeply coupled.

This characteristic is first manifested in the coexistence of the virtual and real combat space. In the virtual space, a parallel space constructed by data streams, AI systems can not only map the situation of the physical world in real time, but also predict the enemy’s movement trajectory, evaluate the effectiveness of different combat plans, and even rehearse combat scenarios that have not yet occurred through deep learning and simulation. This precise mapping and advanced simulation of the physical world by the virtual space greatly expands the dimensions of battlefield perception and decision-making. Secondly, it is manifested in the cross-domain penetration of combat operations. A successful cyberattack may directly paralyze the enemy’s command and control system, turning its real-world tank formations into “steel graveyards”; similarly, the physical destruction of key power grids, communication nodes, etc., will also severely damage the opponent’s combat capabilities in the virtual space. Finally, it is manifested in the increasing blurring of combat subjects and means. With the support of AI technology, cyber hacker organizations and even individuals may become “asymmetric participants” with significant influence; attack methods are also becoming more generalized: a carefully designed piece of malicious code, a data poisoning attack targeting a key industrial control system, or a false intelligence inducement using AI can all trigger a chain reaction in the physical world. Attackers can remain hidden within networks, making them difficult for defenders to trace. This high degree of anonymity and unpredictability significantly increases the cost of defense. Faced with this complex situation, there is an urgent need to transform traditional defensive thinking and combat models, accelerate the construction of a unified confrontation system covering both physical and virtual spaces, develop intelligent combat methods that can penetrate physical and virtual barriers and implement cross-domain collaboration, and forge cross-dimensional suppression capabilities that allow physical destruction and digital suppression to take effect simultaneously.

Feature 4: Accelerated Iterative Evolution – From “No Progress, No Regression” to “Slow Progress, No Regression”

The development of military capabilities in the AI ​​era is a continuous and accelerating process. This is not a linear development of equipment generational replacement in the traditional sense, but rather a self-upgrading of intelligent technology: when relevant intelligent systems undergo millions of self-training sessions in a short period, the speed of technological iteration and offensive-defensive transitions is unprecedentedly accelerated. Yesterday’s impregnable “steel walls” may become “rubble” in the data deluge tomorrow; today’s proud “intelligent advantages” may become “outdated symbols” on a training sandbox the day after tomorrow. From this perspective, the form and logic of warfare in the AI ​​era are shifting from “if you don’t advance, you fall behind” to “if you advance slowly, you fall behind.” However, we must recognize that the more sophisticated the intelligent system, the more it relies on factors such as networks and algorithms. When satellite links are blinded, training databases are contaminated, and cloud computing nodes are destroyed, a highly intelligent war machine may instantly transform from “intelligent” to “incompetent.”

At the practical level, it is essential to construct an agile evolutionary mechanism of “anticipation and adaptation.” We must abandon the static mindset that “advanced equipment is the end goal” and view military capability development as a continuously iterating “living entity.” We must establish a keen network for sensing cutting-edge technologies and, through deep integration into the global innovation ecosystem, capture potential “technological singularities” that could disrupt the rules of warfare at the earliest possible moment. Simultaneously, we must establish a rapid feedback loop that seamlessly connects “research, testing, training, and combat,” ensuring that data from every exercise, simulated confrontation, and even virtual simulation becomes “nutrients” for the self-optimization of intelligent systems such as algorithms and models. This drives the dynamic adjustment of operational concepts, organizational structures, and tactics in near real-time, thereby generating a continuous and rapid endogenous driving force for evolution.

Postscript

In the AI ​​era, the nature of warfare has undergone profound changes—this leap towards intelligence presents both challenges and opportunities. We must embrace technology and accelerate the building of intelligent advantages, but we must also remain clear-headed, recognizing that human political, strategic, and value judgments are always the foundation for navigating future warfare and winning future battlefields. Only by closely integrating the sophistication of technological development with the depth of human wisdom can we consistently remain at the forefront of this wave of military transformation.

Source: 
PLA Daily | May 26, 2026, 05:35

現代國語：

AI正推動戰爭形態進入智慧化時代。傳統戰爭的製勝機理是“大吃小”，比拼的是兵力規模和火力密度；信息化戰爭的製勝機理是“快吃慢”，追求發現即摧毀的打擊效果；而AI戰爭的製勝機理是“精吃粗”——誰能通過算法實現更精細的感知、更精確的判斷、更精確的打擊權，誰就能掌握戰場權。在這種情況下，釐清AI時代的戰爭形態特徵，可以幫助我們更好掌握、運用這項制勝機理，進而在未來戰場占得先機、贏得主動。

特徵一：全局融合感知－從“迷霧重重”到“透明戰場”

AI正在重塑戰場感知模式，以實現前所未有的戰場透明化。其核心在於打破軍種與作戰域間的壁壘，有效整合陸、海、空、天、網、電等多維空間所產生的異質資料流，透過即時融合與處理，建構起一張統一、動態、全景式的戰場態勢圖。在理想狀態下，指揮官「目光所及」的將不再是割裂的局部資訊碎片，而是全域戰場態勢的完整呈現，為其決策提供全維度資訊支撐。值得注意的是，這種透明化高度依賴資料的純度與時效性。掌握關鍵、正確且即時更新數據的一方擁有開啟“透視”戰場的鑰匙；而能有效實施數據污染、注入虛假信息或進行電磁頻譜欺騙的一方，則能使對手陷入更深的“戰場迷霧”，造成其態勢認知的混亂甚至盲目。因此，我們必須清醒認識到，AI帶來的戰場透明化不等於“全知”：人工智能雖能提供近乎完整的“戰場像素”，但每個“像素”的真實性、時效性乃至其背後信息的關聯性，仍需依靠強大的信息驗證與研判機制來確保其準確性。

要做好這一點，需要建立人機深度協同的資訊研究機制。在AI技術的加持下，「人+機器」成為全新作戰編組。 AI是高速運算、不知疲倦的“超級參謀”，人則是洞察全局、把握方向的最終決策者。在這個過程中，AI發揮其快、準、可長時間高強度運行等優勢，在海量數據的初步篩選、關鍵特徵的自動識別、潛在威脅的快速預警等方面“大顯身手”；人則需要發揮主觀能動性，看見AI可能忽視的“戰爭倫理”，守住AI可能忽視的“政治立場”，廓清性，看見AI可能忽視的“戰爭倫理”，守住AI可能忽視的“政治立場”，廓清性，始終為清人的決策，

特徵二：殺傷鏈快速閉合－從“小時”到“毫秒”

AI技術的深度融入，正驅動現代戰爭殺傷鏈閉合速度發生跨越式升級：從機械化時代的“小時級”、信息化戰爭的“分鐘級”，躍升至智能化背景下的“毫秒級”。這項變化源自於演算法對作戰流程的深度重構，傳統線性殺傷鏈演變為動態響應網絡，使AI展現出強大的戰場自主性，能夠自主識別高價值目標、自主規劃最優處置方案、自主分配至最合適的作戰單元執行。這讓「發現即摧毀」的作戰目標不斷趨近現實：目標一旦被感測器捕獲，其座標、屬性、威脅等級判定、最優打擊方案生成、武器平台分配及火力釋放，可在較短時間內同步完成，實現從感知到行動的瞬時閉環。

從這個角度講，演算法即武器——代碼可能比導彈更致命。現代戰爭的較量，已從單一武器平台的表現比拼，躍升至整個作戰體系的智慧化對抗。一套能夠實現毫秒閉環的智慧殺傷鏈系統，其威力不僅在於它能調動多少高精尖武器，更在於它能更自主地完成「決策—行動」循環。在這個過程中，程式碼所建構的自主決策邏輯、資料流驅動的跨域協同網路、算力支撐的即時動態響應，它們共同構成了新型戰鬥力生成的核心要素。此時，針對演算法系統的軟殺傷可能產生戰略級效果，例如，破壞敵決策模型、毀傷敵雲運算節點、斬斷敵戰場資料鏈等，其打擊效果相較於用飛彈摧毀敵重型武器庫可能有過之而無不及。因此，運籌未來戰爭，必須將制演算法權置於與制資訊權等新興制權同樣重要的戰略位置，著力建構具備毫秒自主決策能力與強抗干擾韌性相輔相成的智慧作戰體系，確保在複雜對抗環境中始終牢牢掌控作戰節奏。

特徵三：虛實邊界消失－從“現實戰場”到“全局戰場”

進入資訊化智能化戰爭時代，傳統戰場中涇渭分明的物理與虛擬邊界正在加速消融，逐漸演變為一個物理域、資訊域和社會域高度融合、相互作用的「全局戰場」。實體空間的火力交鋒與數位空間的比特攻防不再是兩條平行線，而是相互交織、彼此塑造、深度耦合的統一整體。

這項特徵首先體現為作戰空間的虛實共生。在虛擬空間這個由資料流建構的平行空間裡，AI系統不僅能即時映射物理世界的態勢，更能透過深度學習和模擬推演，預測敵方的行動軌跡、評估不同作戰方案的效能，甚至預演尚未發生的戰鬥場景。這種虛擬空間對物理世界的精確映射與超前推演，大大拓展了戰場感知與決策的維度。其次，體現為作戰行動的跨域穿透性。一次成功的網路攻擊，可能直接癱瘓敵方的指揮控制系統，導致其現實中的坦克集群成為「鋼鐵墳墓」；同樣，對關鍵電網、通訊節點等的物理摧毀，也將重創對手在虛擬空間的作戰能力。最後，體現為作戰主體與手段的日益模糊化。在AI技術的加持下，網路駭客組織甚至個人，都可能成為具備重大影響力的「非對稱參與者」；攻擊手段同樣泛化：一段精心設計的惡意程式碼、一場針對關鍵工業控制系統的資料投毒、一次利用AI生成的虛假情報誘導，都可能引發物理世界的連鎖反應。攻擊者可以隱匿於網路中，讓防禦者難以溯源。這種高度的匿名性和不可預測性，大大增加了防禦的成本。面對此複雜情況，亟需轉變傳統的防禦思維與作戰模式，加速建構覆蓋實體與虛擬空間的統一對抗體系，發展能穿透虛實屏障、實施跨域協同的智慧作戰手段，鍛造物理摧毀與數位壓制同步生效的跨維壓制能力。

特徵四：迭代進化加速－從“不進則退”到“慢進則退”

AI時代的軍事能力發展，是一個持續的、加速的過程。這並非傳統意義上裝備代際更替的線性發展，而是智能技術的自我升級：當相關智能係統在較短時間內進行百萬次自我訓練，技術迭代、攻防轉換的速度被空前加速，昨日牢不可破的“鋼鐵城牆”，明日可能淪為數據洪流中的“斷壁殘垣”；今天引以為沙盤上的“智能化優勢”，女王上將陳垣”；今天將引以為沙天的“智能化優勢”。從這個角度講，AI時代的戰爭形態和戰爭邏輯正從「不進則退」轉變為「慢進則退」。但我們必須認識到，越是精密的智慧型系統，越是依賴網路、演算法等因素。當衛星連結被致盲、訓練資料庫遭污染、雲端算力節點被摧毀，高度智慧化的戰爭機器可能瞬間從「智慧」變成「無能」。

在實踐層面，必須建構「預見—適應」的敏捷進化機制。摒棄“裝備先進即終點”的靜態思維，將軍事能力建設視為持續迭代的“生命體”，建立敏銳的前沿技術感知網絡，通過深度融入全局創新生態，在第一時間捕捉可能顛覆戰爭規則的“技術奇點”。同時，建立“研、試、訓、戰”無縫銜接的快速反饋閉環，讓每一次演習、模擬對抗甚至虛擬推演的數據，都成為算法模型等智能係統自我優化的“養分”，驅動作戰概念、編制編成、戰術戰法在近乎實時的狀態下動態調整，進而催生出持續、快速進化的內生動力。

編後

AI時代，戰爭形態已發生深刻變化──這場智慧化躍遷，既是挑戰，也是機會。我們既要擁抱技術，加速建構智慧優勢，更要保持清醒，體認到人的政治判斷、戰略判斷、價值判斷等始終是駕馭未來戰爭、制勝未來戰場的基礎。唯有將科技發展之「精」與人類智慧之「深」緊密結合，才能在這場軍事變革浪潮中始終「勇立潮頭」。

中國原創軍事資源：http://www.81.cn/