中國軍事智能化戰爭時代迅速來臨

現代英語：

Since the beginning of the new century, the rapid development of intelligent technologies, with artificial intelligence (AI) at its core, has accelerated the process of a new round of military revolution, and competition in the military field is rapidly moving towards an era of intellectual dominance. Combat elements represented by “AI, cloud, network, cluster, and terminal,” combined in diverse ways, constitute a new battlefield ecosystem, completely altering the mechanisms of victory in warfare. AI systems based on models and algorithms will be the core combat capability, permeating all aspects and stages, playing a multiplicative, transcendent, and proactive role. Platforms are controlled by AI, clusters are guided by AI, and systems are made to decision by AI. Traditional human-centric tactics are being replaced by AI models and algorithms, making intellectual dominance the core control in future warfare. The stronger the intelligent combat capability, the greater the hope of subduing the enemy without fighting.

[Author Biography] Wu Mingxi is the Chief Scientist and Researcher of China Ordnance Industry Group, Deputy Secretary-General of the Science and Technology Committee of China Ordnance Industry Group, and Deputy Director of the Science and Technology Committee of China Ordnance Science Research Institute. His research focuses on national defense science and technology and weaponry development strategies and planning, policies and theories, management and reform research. His major works include “Intelligent Warfare – AI Military Vision,” etc.

Competition in the Age of Intellectual Property

The history of human civilization is a history of understanding and transforming nature, and also a history of understanding and liberating oneself. Through the development of science and technology and the creation and application of tools, humanity has continuously enhanced its capabilities, reduced its burdens, freed itself from constraints, and liberated itself. The control of war has also constantly changed, enriched, and evolved with technological progress, the expansion of human activity space, and the development of the times. Since the 19th century, humanity has successively experienced the control and struggle for land power, sea power, air power, space power, and information power. With the rapid development of intelligent technologies such as artificial intelligence (AI), big data, cloud computing, bio-interdisciplinary technologies, unmanned systems, and parallel simulation, and their deep integration with traditional technologies, humanity’s ability to understand and transform nature has been transformed in terms of epistemology, methodology, and operational mechanisms. This is accelerating the major technological revolutions in machine intelligence, bionic intelligence, swarm intelligence, human-machine integrated intelligence, and intelligent perception, intelligent decision-making, intelligent action, intelligent support, as well as intelligent design, research and development, testing, and manufacturing, thus accelerating the evolution of warfare towards the control and struggle for intellectual power.

The rapid development of intelligent technology has garnered significant attention from major countries worldwide, becoming a powerful driving force for the leapfrog development of military capabilities. The United States and Russia have placed intelligent technology at the core of maintaining their strategic status as global military powers, and significant changes have occurred in their development concepts, models, organizational methods, and innovative applications. They have also carried out substantive applications and practices of military intelligence (see Figure 1).

In August 2017, the U.S. Department of Defense stated that future AI warfare was inevitable and that the U.S. needed to “take immediate action” to accelerate the development of AI warfare technologies. The U.S. military’s “Third Offset Strategy” posits that a military revolution, characterized by intelligent armies, autonomous equipment, and unmanned warfare, is underway; therefore, they have identified intelligent technologies such as autonomous systems, big data analytics, and automation as key development directions. In June 2018, the U.S. Department of Defense announced the establishment of the Joint Artificial Intelligence Center, which, guided by the national AI development strategy, coordinates the planning and construction of the U.S. military’s intelligent military system. In February 2019, then-President Trump signed the “American Artificial Intelligence Initiative” executive order, emphasizing that maintaining U.S. leadership in AI is crucial for safeguarding U.S. economic and national security, and requiring the federal government to invest all resources in promoting innovation in the U.S. AI field. In March 2021, the U.S. National Security Council on Artificial Intelligence released a research report stating that, “For the first time since World War II, the technological advantage that has been the backbone of U.S. economic and military power is under threat. If current trends do not change, China possesses the power, talent, and ambition to surpass the United States as the global leader in artificial intelligence within the next decade.” The report argues that the United States must use artificial intelligence swiftly and responsibly to prepare for these threats in order to safeguard national security and enhance defense capabilities. The report concludes that artificial intelligence will transform the world, and the United States must take a leading role.

Russia also attaches great importance to the technological development and military application of artificial intelligence. The Russian military generally believes that artificial intelligence will trigger the third revolution in the military field, following gunpowder and nuclear weapons. In September 2017, Russian President Vladimir Putin publicly stated that artificial intelligence is the future of Russia, and whoever becomes the leader in this field will dominate the world. In October 2019, Putin approved the “Russian National Strategy for the Development of Artificial Intelligence until 2030,” aiming to accelerate the development and application of artificial intelligence in Russia and seek a world-leading position in the field.

In July 2017, the State Council of China issued the “New Generation Artificial Intelligence Development Plan,” which put forward the guiding ideology, strategic goals, key tasks and safeguard measures for the development of new generation artificial intelligence towards 2030, and deployed efforts to build a first-mover advantage in the development of artificial intelligence and accelerate the construction of an innovative country and a world-class science and technology power.

Other major countries and military powers around the world have also launched their own artificial intelligence development plans, indicating that the global struggle for “intellectual power” has fully unfolded. Land power, sea power, air power, space power, information power, and intellectual power are all results of technological progress and products of their time, each with its own advantages and disadvantages, and some theories are constantly expanding with the changing times. From the development trend of control over warfare since modern times, it can be seen that information power and intellectual power involve the overall situation, carrying greater weight and influence. In the future, with the accelerated pace of intelligent development, intellectual power will become a rapidly growing new type of battlefield control with greater strategic influence on the overall combat situation.

The essence of military intelligence lies in leveraging intelligent technologies to establish diverse identification, decision-making, and control models for the war system. These models constitute artificial intelligence (AI), the core of the new era’s intellectual power struggle. The war system encompasses: equipment systems such as individual units, clusters, manned/unmanned collaborative operations, and multi-domain and cross-domain warfare; combat forces such as individual soldiers, squads, detachments, combined arms units, and theater command; operational links such as networked perception, mission planning and command, force coordination, and comprehensive support; specialized systems such as network attack and defense, electronic warfare, public opinion control, and infrastructure management; and military industrial capabilities such as intelligent design, research and development, production, mobilization, and support. AI, in the form of chips, algorithms, and software, is embedded in every system, level, and link of the war system, forming a systematic brain. Although AI is only a part of the war system, its increasingly powerful “brain-like” functions and capabilities “surpassing human limits” will inevitably dominate the overall situation of future warfare.

Battlefield Ecosystem Reconstruction

Traditional warfare involves relatively independent and separate combat elements, resulting in a relatively simple battlefield ecosystem, primarily consisting of personnel, equipment, and tactics. In the intelligent era, warfare is characterized by significant integration, correlation, and interaction among various combat elements. This will lead to substantial changes in the battlefield ecosystem, forming a combat system, cluster system, and human-machine system comprised of an AI brain, distributed cloud, communication networks, collaborative groups, and various virtual and physical terminals—collectively known as the “AI, Cloud, Network, Cluster, Terminal” intelligent ecosystem (see Figure 2). Among these, AI plays a dominant role.

AI Brain System. The AI ​​brain system of the intelligent battlefield is a networked and distributed system that is inseparable from and interdependent with combat platforms and missions. It can be classified in several ways. Based on function and computing power, it mainly includes cerebellum, swarm brain, midbrain, hybrid brain, and cerebrum; based on combat missions and stages, it mainly includes sensor AI, combat mission planning and decision-making AI, precision strike and controllable destruction AI, network attack and defense AI, electronic warfare AI, intelligent defense AI, and integrated support AI; based on form, it mainly includes embedded AI, cloud AI, and parallel system AI.

The cerebellum mainly refers to the embedded AI in sensor platforms, combat platforms, and support platforms, which mainly performs tasks such as battlefield environment detection, target recognition, rapid maneuver, precision strike, controlled destruction, equipment support, maintenance support, and logistical support.

“Swarm brain” mainly refers to the AI ​​that enables intelligent control of unmanned swarm platforms on the ground, in the air, at sea, in the water, and in space. It mainly performs tasks such as collaborative perception of the battlefield environment, swarm maneuver, swarm attack, and swarm defense. The key components include algorithms for homogeneous swarm systems and algorithms for heterogeneous systems such as manned-unmanned collaboration.

The midbrain mainly refers to the AI ​​system of the command center, data center, and edge computing of the front-line units on the battlefield. It mainly performs dynamic planning, autonomous decision-making, and auxiliary decision-making for tactical unit combat missions under online and offline conditions.

Hybrid brain mainly refers to a hybrid decision-making system in which commanders and machine AI collaborate in combat operations of organized units. Before the battle, it mainly performs human-based combat mission planning; during the battle, it mainly performs adaptive dynamic mission planning and adjustment based on machine AI; and after the battle, it mainly performs hybrid decision-making tasks oriented towards counter-terrorism and defense.

The “brain” primarily refers to the model, algorithm, and tactical libraries of the theater command center and data center, playing a key supporting role in campaign and strategic decision-making. Due to the abundant data, various battlefield AI systems can be trained and modeled here, and then loaded into different mission systems once mature.

In future battlefields, there will be other AIs of different functions, types, and sizes, such as sensor AI, which mainly includes image recognition, electromagnetic spectrum recognition, sound recognition, speech recognition, and human activity behavior recognition. With the rapid development and widespread application of intelligence, AIs of all sizes will exist throughout society, serving the public and society in peacetime, and potentially serving the military in wartime.

Distributed cloud. Military cloud differs from civilian cloud. Generally speaking, a military cloud platform is a distributed resource management system that uses communication networks to search, collect, aggregate, analyze, calculate, store, and distribute operational information and data. By constructing a distributed system and a multi-point fault-tolerant backup mechanism, a military cloud platform possesses powerful intelligence sharing capabilities, data processing capabilities, resilience, and self-healing capabilities. It can provide fixed and mobile, public and private cloud services, achieving “one-point collection, everyone sharing,” greatly reducing information flow links, making command processes flatter and faster, and avoiding redundant and decentralized construction at all levels.

From the perspective of future intelligent warfare needs, military cloud needs to construct at least a four-tiered system: tactical front-end cloud, troop cloud, theater cloud, and strategic cloud. Based on operational elements, it can also be divided into specialized cloud systems such as intelligence cloud, situational awareness cloud, firepower cloud, information warfare cloud, support cloud, and nebula.

1. Front-end cloud primarily refers to computing services provided by units, squads, and platforms, including information perception, target identification, battlefield environment analysis, autonomous and assisted decision-making, and operational process and effect evaluation. The role of front-end cloud is mainly reflected in two aspects. First, it facilitates the sharing and collaboration of computing and storage resources among platforms, and the interactive integration of intelligent combat information. For example, if a platform or terminal is attacked, relevant perception information, damage status, and historical data will be automatically backed up, replaced, and updated through a networked cloud platform, and the relevant information will be uploaded to the higher command post. Second, it provides online information services and intelligent software upgrades for offline terminals.

2. Military cloud primarily refers to the cloud systems built at the battalion and brigade level for operations. Its focus is on providing computing services such as intelligent perception, intelligent decision-making, autonomous action, and intelligent support in response to different threats and environments. The goal of military cloud construction is to establish a networked, automatically backed-up, distributed cloud system connected to multiple links with higher-level units. This system should meet the computing needs of different forces, including reconnaissance and perception, mobile assault, command and control, firepower strikes, and logistical support, as well as the computing needs of various combat missions such as tactical joint operations, manned/unmanned collaboration, and swarm offense and defense.

3. Theater Cloud primarily provides battlefield weather, geographical, electromagnetic, human, and social environmental factors and information data for the entire operational area. It offers comprehensive information on troop deployments, weaponry, movement changes, and combat losses for both sides, as well as relevant information from higher command, friendly forces, and civilian support. Theater Cloud should possess networked, customized, and intelligent information service capabilities. It should interconnect with various operational units through military communication networks (space-based, airborne, ground-based, maritime, and underwater) and civilian communication networks (under secure measures) to ensure efficient, timely, and accurate information services.

4. Strategic cloud is mainly established by a country’s defense system and military command organs. It is primarily based on military information and covers comprehensive information and data related to defense technology, defense industry, mobilization support, economic and social support capabilities, as well as politics, diplomacy, and public opinion. It provides core information, assessments, analyses, and suggestions such as war preparation, operational planning, operational schemes, operational progress, battlefield situation, and battle situation analysis; and provides supporting data such as strategic intelligence, the military strength of adversaries, and war mobilization potential.

The various clouds mentioned above are interconnected, exhibiting both hierarchical and horizontal relationships of collaboration, mutual support, and mutual service. The core tasks of the military cloud platform are twofold: first, to provide data and computing support for building an AI-powered intelligent warfare system; and second, to provide operational information, computing, and data support for various combat personnel and weapon platforms. Furthermore, considering the needs of terminals and group operations, it is necessary to pre-process some cloud computing results, models, and algorithms into intelligent chips and embed them into weapon platforms and group terminals, enabling online upgrades or offline updates.

Communication networks. Military communication and network information constitute a complex super-network system. Since military forces primarily operate in land, sea, air, space, field maneuver, and urban environments, their communication networks encompass strategic and tactical communications, wired and wireless communications, secure communications, and civilian communications. Among these, wireless, mobile, and free-space communication networks are the most crucial components of the military network system, and related integrated electronic information systems are gradually established based on these communication networks.

Military communications in the mechanized era primarily followed the platform, terminal, and user, satisfying specific needs but resulting in numerous silos and extremely poor interconnectivity. In the information age, this situation is beginning to change. Currently, military communication networks are adopting new technological systems and development models, characterized by two main features: first, “network-data separation,” where information transmission does not depend on any specific network transmission method—”network access is all that matters”—any information can be delivered as long as the network link is unobstructed; second, internet-based architecture, utilizing IP addresses, routers, and servers to achieve “all roads lead to Beijing,” i.e., military networking or grid-based systems. Of course, military communication networks differ from civilian networks. Strategic and specialized communication needs exist at all times, such as nuclear button communications for nuclear weapons and command and control of strategic weapons, information transmission for satellite reconnaissance, remote sensing, and strategic early warning, and even specialized communications in individual soldier rooms and special operations conditions. These may still adopt a mission-driven communication model. Even so, standardization and internet connectivity are undoubtedly the future trends in military communication network development. Otherwise, not only will the number of battlefield communication frequency bands, radios, and information exchange methods increase, leading to self-interference, mutual interference, and electromagnetic compatibility difficulties, but radio spectrum management will also become increasingly complex. More importantly, it will be difficult for platform users to achieve automatic communication based on IP addresses and routing structures, unlike email on the internet where a single command can be sent to multiple users. Future combat platforms will certainly be both communication user terminals and also function as routers and servers.

Military communication network systems mainly include space-based communication networks, military mobile communication networks, data links, new communication networks, and civilian communication networks.

1. Space-Based Information Networks. The United States leads in the construction and utilization of space-based information networks. This is because more than half of the thousands of orbiting platforms and payloads in space are American-owned. Following the Gulf War, and especially during the Iraq War, the US military accelerated the application and advancement of space-based information networks through wartime experience. After the Iraq War, through the utilization of space-based information and the establishment of IP-based interconnection, nearly 140 vertical “chimneys” from the Gulf War period were completely interconnected horizontally, significantly shortening the “Out-of-Target-Action” (OODA) loop time. The time from space-based sensors to the shooter has been reduced from tens of hours during the Gulf War to approximately 20 seconds currently using artificial intelligence for identification.

With the rapid development of small satellite technology, low-cost, multi-functional small satellites are becoming increasingly common. As competition intensifies in commercial launches, costs are dropping dramatically, and a single launch can carry several, a dozen, or even dozens of small satellites. If miniaturized electronic reconnaissance, visible light and infrared imaging, and even quantum dot micro-spectroscopy instruments are integrated onto these satellites, achieving integrated reconnaissance, communication, navigation, meteorological, and mapping functions, the future world and battlefield will become much more transparent.

2. Military Mobile Communication Networks. Military mobile communication networks have three main uses. First, command and control between various branches of the armed forces and combat units in joint operations; this type of communication requires a high level of confidentiality, reliability, and security. Second, communication between platforms and clusters, requiring anti-jamming capabilities and high reliability. Third, command and control of weapon systems, mostly handled through data links.

Traditional military mobile communication networks are mostly “centralized, vertically focused, and tree-like structures.” With the acceleration of informatization, the trend towards “decentralized, self-organizing networks, and internet-based” is becoming increasingly apparent. As cognitive radio technology matures and is widely adopted (see Figure 3), future network communication systems will be able to automatically identify electromagnetic interference and communication obstacles on the battlefield, quickly locate available spectrum resources, and conduct real-time communication through frequency hopping and other methods. Simultaneously, software and cognitive radio technology can be compatible with different communication frequency bands and waveforms, facilitating seamless transitions from older to newer systems.

3. Data Links. A data link is a specialized communication technology that uses time division, frequency division, and code division to transmit pre-agreed, periodic, or irregular, regular or irregular critical information between various combat platforms. Unless fully understood or deciphered by the enemy, it is very difficult to interfere with. Data links are mainly divided into two categories: dedicated and general-purpose. Joint operations, formation coordination, and swarm operations primarily utilize general-purpose data links. Satellite data links, UAV data links, missile-borne data links, and weapon fire control data links are currently mostly dedicated. In the future, generalization will be a trend, and specialization will decrease. Furthermore, from the perspective of the relationship between platforms and communication, the information transmission and reception of platform sensors and internal information processing generally follow the mission system, exhibiting strong specialization characteristics, while communication and data transmission between platforms are becoming increasingly general-purpose.

4. New Communication Technologies. Traditional military communication primarily relies on microwave communication. Due to its large divergence angle and numerous application platforms, corresponding electronic jamming and microwave attack methods have developed rapidly, making it easy to carry out long-range interference and damage. Therefore, new communication technologies such as millimeter waves, terahertz waves, laser communication, and free-space optical communication have become important choices that are both anti-jamming and easy to implement high-speed, high-capacity, and high-bandwidth communication. Although high-frequency electromagnetic waves have good anti-jamming performance due to their smaller divergence angle, achieving precise point-to-point aiming and omnidirectional communication still presents certain challenges, especially under conditions of high-speed maneuvering and rapid trajectory changes of combat platforms. How to achieve alignment and omnidirectional communication is still under technological exploration.

5. Civilian Communication Resources. The effective utilization of civilian communication resources is a strategic issue that must be considered and cannot be avoided in the era of intelligentization. In the future, leveraging civilian communication networks, especially 5G/6G mobile communications, for open-source information mining and data correlation analysis to provide battlefield environment, target, and situational information will be crucial for both combat and non-combat military operations. In non-combat military operations, especially overseas peacekeeping, rescue, counter-terrorism, and disaster relief, the military’s dedicated communication networks can only be used within limited areas and regions, raising the question of how to communicate and connect with the outside world. There are two main ways to utilize civilian communication resources: one is to utilize civilian satellite communication resources, especially small satellite communication resources; the other is to utilize civilian mobile communication and internet resources.

The core issue in the interactive utilization of military and civilian communication resources is addressing security and confidentiality. One approach is to employ firewalls and encryption, directly utilizing civilian satellite communications and global mobile communication infrastructure for command and communication; however, the risks of hacking and cyberattacks remain. Another approach is to utilize emerging technologies such as virtualization, intranets, semi-physical isolation, one-way transmission, mimicry defense, and blockchain to address these challenges.

Collaborative swarms. By simulating the behavior of bee colonies, ant colonies, flocks of birds, and schools of fish in nature, this research studies the autonomous collaborative mechanisms of swarm systems such as drones and smart munitions to accomplish combat missions such as attacking or defending against enemy targets. This can achieve strike effects that are difficult to achieve with traditional combat methods and approaches. Collaborative swarms are an inevitable trend in intelligent development and a major direction and key area of ​​intelligent construction. No matter how advanced the combat performance or how powerful the functions of a single combat platform, it cannot form a collective or scalable advantage. Simply accumulating quantity and expanding scale, without autonomous, collaborative, and orderly intelligent elements, is just a disorganized mess.

Collaborative swarms mainly comprise three aspects: first, manned/unmanned collaborative swarms formed by the intelligent transformation of existing platforms, primarily constructed from large and medium-sized combat platforms; second, low-cost, homogeneous, single-function, and diverse combat swarms, primarily constructed from small unmanned combat platforms and munitions; and third, biomimetic swarms integrating human and machine intelligence, possessing both biological and machine intelligence, primarily constructed from highly autonomous humanoid, reptile-like, avian-like, and marine-like organisms. Utilizing collaborative swarm systems for cluster warfare, especially swarm warfare, offers numerous advantages and characteristics.

1. Scale Advantage. A large unmanned system can disperse combat forces, increasing the number of targets the enemy can attack and forcing them to expend more weapons and ammunition. The survivability of a swarm, due to its sheer number, is highly resilient and resilient; the survivability of a single platform becomes less important, while the overall advantage becomes more pronounced. The sheer scale prevents drastic fluctuations in combat effectiveness, because unlike high-value manned combat platforms and complex weapon systems such as the B-2 strategic bomber and advanced F-22 and F-35 fighter jets, the loss of a low-cost unmanned platform, once attacked or destroyed, results in a sharp decline in combat effectiveness. Swarm operations can launch simultaneous attacks, overwhelming enemy defenses. Most defensive systems have limited capabilities, able to handle only a limited number of threats at a time. Even with dense artillery defenses, a single salvo can only hit a limited number of targets, leaving some to escape. Therefore, swarm systems possess extremely strong penetration capabilities.

2. Cost Advantage. Swarm warfare, especially bee warfare, primarily utilizes small and medium-sized UAVs, unmanned platforms, and munitions. These have simple product lines, are produced in large quantities, and have consistent quality and performance requirements, facilitating low-cost mass production. While the pace of upgrades and replacements for modern weapons and combat platforms has accelerated significantly, the cost increases have also been staggering. Since World War II, weapons development and procurement prices have shown that equipment costs and prices have risen much faster than performance improvements. Main battle tanks during the Gulf War cost 40 times more than those during World War II, while combat aircraft and aircraft carriers cost as much as 500 times more. From the Gulf War to 2020, the prices of various main battle weapons and equipment increased several times, tens of times, or even hundreds of times. In comparison, small and medium-sized UAVs, unmanned platforms, and munitions with simple product lines have a clear cost advantage.

3. Autonomous Advantage. Under a unified spatiotemporal reference platform, through networked active and passive communication and intelligent perception of battlefield targets, individual platforms in the group can accurately perceive the distance, speed, and positional relationships between each other. They can also quickly identify the nature, size, priority, and distance of target threats, as well as their own distance from neighboring platforms. With pre-defined operational rules, one or more platforms can conduct simultaneous or wave-based attacks according to the priority of target threats, or they can attack in groups simultaneously or in multiple waves (see Figure 4). Furthermore, the priority order for subsequent platforms to replace a damaged platform can be clearly defined, ultimately achieving autonomous decision-making and action according to pre-agreed operational rules. This intelligent combat operation, depending on the level of human involvement and the difficulty of controlling key nodes, can be either completely autonomous, or semi-autonomous, with human intervention.

4. Decision-making advantage. The future battlefield environment is becoming increasingly complex, with combatants vying for dominance in intense strategic maneuvering and confrontation. Therefore, relying on humans to make decisions in a high-intensity confrontation environment is neither timely nor reliable. Thus, only by entrusting automated environmental adaptation, automatic target and threat identification, autonomous decision-making, and coordinated action to collaborative groups can adversaries be rapidly attacked or effective defenses implemented, thereby gaining battlefield advantage and initiative.

The coordination group brings new challenges to command and control. How to implement command and control of the cluster is a new strategic issue. Control can be implemented in a hierarchical and task-based manner, which can be roughly divided into centralized control mode, hierarchical control mode, consistent coordination mode, and spontaneous coordination mode. [1] Various forms can be adopted to achieve human control and participation. Generally speaking, the smaller the tactical unit, the more autonomous action and unmanned intervention should be adopted; at the level of organized unit operations, since the control of multiple combat groups is involved, centralized planning and hierarchical control are required, and human participation should be limited; at the higher strategic and operational levels, the cluster is only used as a platform weapon and combat style, which requires unified planning and layout, and the degree of human participation will be higher. From the perspective of mission nature, the operation and use of strategic weapons, such as nuclear counterattacks, requires human operation and is not suitable for autonomous handling by weapon systems. When conducting offensive and defensive operations against important or high-value targets, such as decapitation strikes, full human participation and control are necessary, while simultaneously leveraging the autonomous functions of the weapon systems. For offensive operations against tactical targets, if the mission requires lethal strikes and destruction, limited human participation is permissible, or, after human confirmation, the coordinated group can execute the operation automatically. When performing non-strike missions such as reconnaissance, surveillance, target identification, and clearance, or short-duration missions such as air defense and missile defense where human involvement is difficult, the coordinated group should primarily execute these tasks automatically, without human involvement. Furthermore, countermeasures for swarm operations must be carefully studied. Key research should focus on countermeasures against electronic deception, electromagnetic interference, cyberattacks, and high-power microwave weapons, electromagnetic pulse bombs, and artillery-missile systems, as their effects are relatively significant. Simultaneously, research should be conducted on countermeasures such as laser weapons and swarm-to-swarm tactics, gradually establishing a “firewall” that humans can effectively control against coordinated groups.

Virtual and physical terminals. Virtual and physical terminals mainly refer to various terminals linked to the cloud and network, including sensors with pre-embedded intelligent modules, command and control platforms, weapon platforms, support platforms, related equipment and facilities, and combat personnel. Future equipment and platforms will be cyber-physical systems (CPS) and human-computer interaction systems with diverse front-end functions, cloud-based back-end support, virtual-physical interaction, and online-offline integration. Simple environmental perception, path planning, platform maneuverability, and weapon operation will primarily rely on front-end intelligence such as bionic intelligence and machine intelligence. Complex battlefield target identification, combat mission planning, networked collaborative strikes, combat situation analysis, and advanced human-computer interaction will require information, data, and algorithm support from back-end cloud platforms and cloud-based AI. The front-end intelligence and back-end cloud intelligence of each equipment platform should be combined for unified planning and design, forming a comprehensive advantage of integrated front-end and back-end intelligence. Simultaneously, virtual soldiers, virtual staff officers, virtual commanders, and their intelligent and efficient interaction with humans are also key areas and challenges for future research and development.

Qualitative change in the form of warfare

Since modern times, human society has mainly experienced large-scale mechanized warfare and smaller-scale informationized local wars. The two world wars that occurred in the first half of the 20th century were typical examples of mechanized warfare. The Gulf War, the Kosovo War, the Afghanistan War, the Iraq War, and the Syrian War since the 1990s fully demonstrate the form and characteristics of informationized warfare. In the new century and new stage, with the rapid development and widespread application of intelligent technologies, the era of intelligent warfare, characterized by data and computing, models and algorithms, is about to arrive (see Figure 5).

Mechanization is a product of the industrial age, focusing on mechanical power and electrical technology. Its weaponry primarily manifests as tanks, armored vehicles, artillery, aircraft, and ships, corresponding to mechanized warfare. Mechanized warfare is mainly based on classical physics, represented by Newton’s laws, and large-scale socialized production. It is characterized by large-scale, linear, and contact warfare. Tactically, it typically involves on-site reconnaissance, terrain surveys, understanding the opponent’s forward and rear deployments, making decisions based on one’s own capabilities, implementing offensive or defensive maneuvers, and assigning tasks, coordinating operations, and ensuring logistical support. It exhibits clear characteristics such as hierarchical command and control and sequential temporal and spatial operations.

Information technology, a product of the information age, focuses on information technologies such as computers and network communications. Its equipment primarily manifests as radar, radios, satellites, missiles, computers, military software, command and control systems, cyber and electronic warfare systems, and integrated electronic information systems, corresponding to the form of information warfare. Information warfare is mainly based on the three laws of computers and networks (Moore’s Law, Gilder’s Law, and Metcalfe’s Law), emphasizing integrated, precise, and three-dimensional operations. It establishes a seamless and rapid information link from sensor to shooter, seizing information dominance and achieving preemptive detection and strike. Tactically, it requires detailed identification and cataloging of the battlefield and targets, highlighting the role of networked perception and command and control systems, and placing new demands on the interconnectivity and other information functions of platforms. Due to the development of global information systems and diversified network communications, information warfare blurs the lines between front and rear lines, emphasizing horizontal integration of reconnaissance, control, strike, assessment, and support, as well as the integration and flattening of strategy, campaign, and tactics.

Intelligentization is a product of the knowledge economy era. Technologically, it focuses on intelligent technologies such as artificial intelligence, big data, cloud computing, cognitive communication, the Internet of Things, biological cross-disciplinary, hybrid enhancement, swarm intelligence, autonomous navigation and collaboration. In terms of equipment, it mainly manifests as unmanned platforms, intelligent munitions, swarm systems, intelligent sensing and database systems, adaptive mission planning and decision-making systems, combat simulation and parallel training systems, military cloud platforms and service systems, public opinion early warning and guidance systems, and intelligent wearable systems, which correspond to the form of intelligent warfare.

Intelligent warfare, primarily based on biomimetic, brain-like principles, and AI-driven battlefield ecosystems, is a new combat form characterized by “energy mobility and information interconnection,” supported by “network communication and distributed cloud,” centered on “data computing and model algorithms,” and focused on “cognitive confrontation.” It features multi-domain integration, cross-domain offense and defense, unmanned operation, cluster confrontation, and integrated interaction between virtual and physical spaces.

Intelligent warfare aims to meet the needs of nuclear and conventional deterrence, joint operations, all-domain operations, and non-war military operations. It focuses on multi-domain integrated operations encompassing cognitive, informational, physical, social, and biological domains, exhibiting characteristics such as distributed deployment, networked links, flattened structures, modular combinations, adaptive reconfiguration, parallel interaction, focused energy release, and nonlinear effects. Its winning mechanisms overturn traditions, its organizational forms undergo qualitative changes, its operational efficiency is unprecedentedly improved, and its combat power generation mechanisms are transformed. These substantial changes are mainly reflected in the following ten aspects.

The Winning Mechanism Dominated by AI. Under intelligent conditions, new combat elements represented by “AI, cloud, network, cluster, and terminal” will reshape the battlefield ecosystem, completely changing the winning mechanism of war. Among them, AI systems based on models and algorithms are the core combat capability, permeating all aspects and links, playing a multiplicative, transcendent, and proactive role. Platforms are controlled by AI, clusters are guided by AI, and systems are made by AI. The traditional human-based combat methods are being replaced by AI models and algorithms. Algorithmic warfare will play a decisive role in war, and the combat system and process will ultimately be dominated by AI. The right to intelligence will become the core control in future warfare.

Different eras and different forms of warfare result in different battlefield ecosystems, with entirely different compositions of combat elements and winning mechanisms. Mechanized warfare is platform-centric warfare, with “movement” as its core and firepower and mobility as its dominant forces, pursuing energy delivery and release through equipment. Combat elements mainly include: personnel + mechanized equipment + tactics. The winning mechanism is based on human-led decision-making in the operational use of mechanized equipment, achieving victory with superior numbers, overwhelming smaller forces, and controlling slower forces, with comprehensive, efficient, and sustainable mobilization capabilities playing decisive or important roles. Informationized warfare is network-centric warfare, with “connectivity” as its core and information power as its dominant force, pursuing energy aggregation and release through networks. Combat elements and their interrelationships mainly consist of “personnel + informationized equipment + tactics” based on network information. Information permeates personnel, equipment, and tactics, establishing seamless information connections “from sensor to shooter,” achieving system-wide and networked combat capabilities, using systems against localized forces, networks against discrete forces, and speed against slow forces, becoming a crucial mechanism for achieving victory in war. Information plays a multiplier role in equipment and combat systems, but the platform remains human-centric. Information assists in decision-making, but most decisions are still made by humans. Intelligent warfare is cognitive-centric warfare, with “computation” at its core and intelligence as the dominant force. Intelligence will carry more weight than firepower, mobility, and information power, pursuing the use of intelligence to control and dominate capabilities, using the virtual to overcome the real, and achieving victory through superiority. The side with more AI and whose AI is smarter will have greater initiative on the battlefield. The main combat elements and their interrelationships are: AI × (cloud + network + swarm + human + equipment + tactics), which can be simplified to an interconnected and integrated battlefield ecosystem composed of “AI, cloud, network, swarm, and terminal” elements. In the future, AI’s role in warfare will become increasingly significant and powerful, ultimately playing a decisive and dominant role.

Emphasizing the leading role of AI does not deny the role of humans in warfare. On the one hand, human intelligence has been pre-emptively utilized and endowed into AI; on the other hand, at the pre-war, post-war, and strategic levels, for a considerable period of time and in the foreseeable future, AI cannot replace humans.

Modern warfare is becoming increasingly complex, with combat operations moving at ever faster paces. The ability to quickly identify and process massive amounts of information, respond rapidly to battlefield situations, and formulate decisive strategies is far beyond human capability and exceeds the limits of current technology (see Tables 1 and 2). As AI becomes more widely applied and plays a more significant role in warfare, operational processes will be reshaped, and the military kill chain will be accelerated and made more efficient. Rapid perception, decision-making, action, and support will become crucial factors for victory in future intelligent warfare.

In the future, intelligent recognition and pattern recognition of images, videos, electromagnetic spectrum, and voice will enable rapid and accurate target identification from complex battlefield information gathered by air, land, and sea sensor networks. Utilizing big data technology, through multi-source, multi-dimensional directional search and intelligent correlation analysis, not only can various targets be accurately located, but also human behavior, social activities, military operations, and public opinion trends can be precisely modeled, gradually improving the accuracy of early warning and prediction. Based on precise battlefield information, each theater and battlefield can adaptively implement mission planning, autonomous decision-making, and operational process control through extensive parallel modeling and simulation training in virtual space. AI on various combat platforms and cluster systems can autonomously and collaboratively execute tasks around operational objectives according to mission planning, and proactively adjust to changes that may occur at any time. By establishing a distributed, networked, intelligent, and multi-modal support system and pre-positioned deployment, rapid and precise logistics distribution, material supply, and intelligent maintenance can be implemented. In summary, through the widespread application of intelligent technologies and the proactive and evolving capabilities of various AI systems, the entire operational process—including planning, prediction, perception, decision-making, implementation, control, and support—can be re-engineered to achieve a “simple, fast, efficient, and controllable” operational workflow. This will gradually free humanity from the burdens of arduous combat tasks. Operational workflow re-engineering will accelerate the pace, compress time, and shorten processes on the future battlefield.

The winning mechanism dominated by AI is mainly manifested in combat capabilities, methods, strategies, and measures. It fully integrates human intelligence, approaches human intelligence, surpasses human limits, leverages the advantages of machines, and embodies advancement, disruption, and innovation. This advancement and innovation is not a simple extension or increase in quantity in previous wars, but a qualitative change and leap, a higher-level characteristic. This higher-level characteristic is reflected in intelligent warfare possessing “brain-like” functions and many “capabilities that surpass human limits” that traditional warfare lacks. As AI continues to optimize and iterate, it will one day surpass ordinary soldiers, staff officers, commanders, and even elite and expert groups, becoming a “super brain” and a “super brain group.” This is the core and key of intelligent warfare, a technological revolution in the fields of epistemology and methodology, and a high-level combat capability that humanity can currently foresee, achieve, and evolve.

The role of cyberspace is rising. With the progress of the times and the development of technology, the operational space has gradually expanded from physical space to virtual space. The role and importance of virtual space in the operational system are gradually rising and becoming increasingly important, and it is increasingly deeply integrated with physical space and other fields. Virtual space is an information space based on network electromagnetics constructed by humans. It can reflect human society and the material world from multiple perspectives, and can be utilized by transcending many limitations of the objective world. It is constructed by the information domain, connected by the physical domain, reflected by the social domain, and utilized by the cognitive domain. In a narrow sense, virtual space mainly refers to the civilian Internet; in a broad sense, virtual space mainly refers to cyberspace, including various Internet of Things, military networks, and dedicated networks. Cyberspace is characterized by being easy to attack but difficult to defend, using software to fight hard, integrating peacetime and wartime, and blurring the lines between military and civilian sectors. It has become an important battlefield for conducting military operations, strategic deterrence, and cognitive confrontation.

The importance of cyberspace is mainly reflected in three aspects: First, through network information systems, it connects dispersed combat forces and elements into a whole, forming a systematic and networked combat capability, which becomes the foundation of information warfare; second, it becomes the main battlefield and basic support for cognitive confrontation such as cyberspace, intelligence, public opinion, psychology, and consciousness; and third, it establishes virtual battlefields, conducts combat experiments, realizes virtual-real interaction, and forms the core and key to parallel operations and the ability to use the virtual to defeat the real.

In the future, with the accelerated upgrading of global interconnection and the Internet of Things, and with the establishment, improvement and widespread application of systems such as space-based networked reconnaissance, communication, navigation, mobile internet, Wi-Fi, high-precision global spatiotemporal reference platforms, digital maps, and industry big data, human society and global military activities will become increasingly “transparent,” increasingly networked, perceived, analyzed, correlated, and controlled (see Figure 6). This will have a profound, all-round, and ubiquitous impact on military construction and operations. The combat system in the intelligent era will gradually expand from closed to open, and from military-led to a “source-open and ubiquitous” direction that integrates military and civilian sectors.

In the era of intelligentization, information and data from the physical, informational, cognitive, social, and biological fields will gradually flow freely. Combat elements will achieve deep interconnection and the Internet of Things. Various combat systems will evolve from basic “capability combinations” to advanced “information fusion, data linking, and integrated behavioral interaction,” possessing powerful all-domain perception, multi-domain fusion, and cross-domain combat capabilities, and the ability to effectively control important targets, sensitive groups, and critical infrastructure anytime, anywhere. A report from the U.S. Army Joint Arms Center argues that the world is entering an era of “ubiquitous global surveillance.” Even if the world cannot track all activities, the proliferation of technology will undoubtedly cause the potential sources of information to grow exponentially.

Currently, network-based software attacks have acquired the capability to cause physical damage, and cyberattacks by militarily advanced countries possess operational capabilities such as intrusion, deception, interference, and sabotage. Cyberspace has become another important battlefield for military operations and strategic deterrence. The United States has already used cyberattacks in actual combat. Ben Ali of Tunisia, Gaddafi of Libya, and Saddam Hussein of Iraq were all influenced by US cyberattacks and WikiLeaks, causing shifts in public opinion, psychological breakdowns, and social unrest, leading to the rapid collapse of their regimes and having a disruptive impact on traditional warfare. Through the Snowden revelations, a list of 49 cyber reconnaissance projects across 11 categories used by the United States was gradually exposed. Incidents such as the Stuxnet virus’s sabotage of Iranian nuclear facilities, the Gauss virus’s mass intrusion into Middle Eastern countries, and the Cuban Twitter account’s control of public opinion demonstrate that the United States possesses powerful monitoring capabilities, as well as soft and hard attack and psychological warfare capabilities over the internet, closed networks, and mobile wireless networks.

The war began with virtual space experiments. The US military began exploring combat simulation, operational experiments, and simulation training in the 1980s. Later, the US military pioneered the use of virtual reality, wargaming, and digital twin technologies in virtual battlefields and combat experiments. Analysis shows that the US military conducted combat simulations in military operations such as the Gulf War, the Kosovo War, the Afghanistan War, and the Iraq War, striving to find the optimal operational and action plans. It has been reported that before Russia intervened militarily in Syria, it conducted pre-war exercises in its war labs. Based on the experimental simulations, it formulated the “Center-2015” strategic exercise plan, practicing “mobility and accessibility in unfamiliar areas” for combat in Syria. After the exercise, Russian Chief of the General Staff Gerasimov emphasized that the primary means would be political, economic, and psychological warfare, supplemented by long-range precision air strikes and special operations, ultimately achieving political and strategic objectives. Practice shows that the process of Russia’s intervention in Syria was largely consistent with these experiments and exercises.

In the future, with the application and development of virtual simulation, mixed reality, big data, and intelligent software, a parallel military artificial system can be established, allowing physical forces in the physical space to map and iterate with virtual forces in the virtual space. This will enable rapid, high-intensity adversarial training and supercomputing that are difficult to achieve in the physical space. It can also engage in combat and games against highly realistic “blue force systems,” continuously accumulating data, building models and algorithms, and ultimately using the optimal solutions to guide the construction and combat of physical forces, achieving the goal of virtual-real interaction, using the virtual to control the real, and winning with the virtual. On January 25, 2019, DeepMind, Google’s AI team, and Blizzard Entertainment, the developer of StarCraft, announced the results of the December 2018 match between AlphaSTAR and professional players TLO and MANA. In the best-of-five series, AlphaSTAR won both matches 5-0. AlphaSTAR completed the training workload that would take human players 200 years in just two weeks, demonstrating the enormous advantages and bright prospects of simulated adversarial training in virtual space.

The combat style is dominated by unmanned operations. In the era of intelligentization, unmanned warfare will become the basic form, and the integration and development of artificial intelligence and related technologies will gradually push this form to an advanced stage. Unmanned systems represent the full pre-positioning of human intelligence in the combat system and are a concentrated manifestation of the integrated development of intelligence, informatization, and mechanization. Unmanned equipment first appeared in the field of drones. In 1917, Britain built the world’s first drone, but it was not used in actual combat. With the development of technology, drones were gradually used in target drones, reconnaissance, and reconnaissance-strike integrated operations. Since the beginning of the 21st century, unmanned technologies and equipment have achieved tremendous leaps and major breakthroughs in exploration and application due to their advantages such as mission-centric design, no need to consider crew requirements, and high cost-effectiveness. They have shown a rapid and comprehensive development trend, and their application scope has expanded rapidly, covering various fields such as air, surface, underwater, ground, and space.

In recent years, technologies such as artificial intelligence, bionic intelligence, human-machine integrated intelligence, and swarm intelligence have developed rapidly. With the help of satellite communication and navigation, and autonomous navigation, unmanned combat platforms can effectively achieve remote control, formation flight, and swarm collaboration. Currently, unmanned combat aerial vehicles, underwater unmanned platforms, and space-based unmanned autonomous robots have emerged one after another. Bipedal, quadrupedal, multi-legged, and cloud-based intelligent robots are developing rapidly and have entered the fast lane of engineering and practical application, with military applications not far off.

Overall, unmanned warfare in the era of intelligentization will enter three stages of development. The first stage is the initial stage, characterized by manned dominance and unmanned support, where “unmanned warfare under manned leadership” means that combat behavior is completely controlled and dominated by humans before, during, and after the operation. The second stage is the intermediate stage, characterized by manned support and unmanned dominance, where “unmanned warfare under limited control” means that human control is limited, auxiliary, but crucial throughout the entire combat process, and in most cases, the autonomous action capabilities of the platform can be relied upon. The third stage is the advanced stage, characterized by manned rules and unmanned action, where “unmanned warfare with manned design and minimal control” means that humans conduct overall design in advance, clarifying autonomous behavior and rules of the game under various combat environments, and the execution phase is mainly entrusted to unmanned platforms and unmanned forces for autonomous execution.

Autonomous behavior or autonomy is the essence of unmanned warfare and a common and prominent feature of intelligent warfare, manifested in many aspects.

First, the autonomy of combat platforms, mainly including the autonomous capabilities and intelligence level of unmanned aerial vehicles, ground unmanned platforms, precision-guided weapons, underwater and space robots.

Second, the detection system is autonomous, which mainly includes automatic search, tracking, association, aiming, and intelligent recognition of information such as images, voice, video, and electronic signals.

Thirdly, there is autonomous decision-making, the core of which is AI-based autonomous decision-making within the combat system. This mainly includes automatic analysis of the battlefield situation, automatic planning of combat missions, automated command and control, and intelligent human-machine interaction.

Fourthly, autonomous coordination in combat operations, which initially includes autonomous coordination between manned and unmanned systems, and later includes autonomous unmanned swarms, such as various combat formations, bee swarms, ant swarms, fish swarms, and other combat behaviors.

Fifth, autonomous network attack and defense behaviors, including automatic identification, automatic tracing, automatic protection, and autonomous counterattack against various viruses and network attacks.

Sixth, cognitive electronic warfare, which automatically identifies the power, frequency band, and direction of electronic interference, automatically hops frequencies and autonomously forms networks, and engages in active and automatic electronic interference against adversaries.

Seventh, other autonomous behaviors, including intelligent diagnosis, automatic repair, and self-protection.

In the future, with the continuous upgrading of the integration and development of artificial intelligence and related technologies, unmanned operations will rapidly develop towards autonomy, biomimicry, swarming, and distributed collaboration, gradually pushing unmanned warfare to an advanced stage and significantly reducing direct confrontation between human forces on the battlefield. Although manned platforms will continue to exist in the future, biomimetic robots, humanoid robots, swarm weapons, robot armies, and unmanned system warfare will become the norm in the intelligent era. Since unmanned systems can replace human beings in many combat domains and can accomplish tasks autonomously, unmanned combat systems will always be there to protect humans before they suffer physical attacks or injuries. Therefore, unmanned combat systems in the intelligent era are humanity’s main protective barrier, its shield and shield.

All-domain operations and cross-domain offense and defense. In the era of intelligent warfare, all-domain operations and cross-domain offense and defense are also a fundamental style of combat, manifested in many combat scenarios and aspects. From land, sea, air, and space to multiple domains including physical, information, cognitive, social, and biological domains, as well as the integration and interaction of virtual and physical elements, from peacetime strategic deterrence to wartime high-confrontation, high-dynamic, and high-response operations, the time and space span is enormous. It involves not only physical space operations and cyberspace cyber offense and defense, information warfare, public opinion guidance, and psychological warfare, but also tasks such as global security governance, regional security cooperation, counter-terrorism, and rescue, and the control of critical infrastructure such as networks, communications, power, transportation, finance, and logistics.

Since 2010, supported by advancements in information and intelligent technologies, the U.S. military has proposed concepts such as operational cloud, distributed lethality, multi-domain warfare, algorithmic warfare, mosaic warfare, and joint all-domain operations. The aim is to maintain battlefield and military superiority by using system-wide systems against localized ones, multi-functional systems against simpler ones, multi-domain systems against single-domain ones, integrated systems against discrete ones, and intelligent systems against non-intelligent ones. The U.S. military proposed the concept of multi-domain warfare in 2016 and joint all-domain operations in 2020, aiming to develop cross-service and cross-domain joint operational capabilities, ensuring that each service’s operations are supported by all three services, and possessing all-domain capabilities against multi-domain and single-domain ones.

In the future, with breakthroughs in key technologies for the cross-disciplinary integration of artificial intelligence and multidisciplinary collaboration, multi-domain integration and cross-domain offense and defense based on AI and human-machine hybrid intelligence will become a distinctive feature of intelligent warfare. This will be achieved across functional domains such as physics, information, cognition, society, and biology, as well as geographical domains such as land, sea, air, and space.

In the intelligent era, multi-domain and cross-domain operations will expand from mission planning, physical collaboration, and loose coordination to heterogeneous integration, data linking, tactical interoperability, and cross-domain offensive and defensive integration.

First, multi-domain integration. Based on different battlefields and adversaries in a multi-domain environment, different combat styles, combat procedures and missions are planned in accordance with the requirements of joint operations, and unified as much as possible. This achieves the overall planning and integration of information, firepower, defense, support and command and control, and the integration of combat capabilities at the strategic, operational and tactical levels, forming the capability of one-domain operations and multi-domain joint rapid support.

Second, cross-domain offense and defense. Supported by a unified network information system, and through a unified battlefield situation and data information exchange based on unified standards, the information links for cross-domain joint operations reconnaissance, control, strike, and assessment are completely opened up, enabling seamless integration of operational elements and capabilities at the tactical and fire control levels, as well as collaborative actions between services, cross-domain command and interoperability.

Third, the entire process is interconnected. Multi-domain integration and cross-domain offense and defense are treated as a whole, with coordinated design and interconnectedness throughout. Before the war, intelligence gathering and analysis are conducted, along with public opinion warfare, psychological warfare, propaganda warfare, and necessary cyber and electronic warfare attacks. During the war, special operations and cross-domain actions are used to carry out decapitation strikes, key point raids, and precise and controllable strikes (see Figure 7). After the war, defense against cyberattacks on information systems, elimination of negative public opinion’s impact on the public, and prevention of enemy damage to infrastructure are addressed through post-war governance, public opinion control, and the restoration of social order across multiple areas.

Fourth, AI support. Through combat experiments, simulation training, and necessary test verification and real-world testing, we continuously accumulate data, optimize models, and establish AI combat models and algorithms for different combat styles and adversaries, forming an intelligent brain system to better support joint operations, multi-domain operations, and cross-domain offense and defense.

Human-AI hybrid decision-making. The continuous improvement, optimization, upgrading, and perfection of the AI ​​brain system in intelligent battlefields will enable it to surpass humans in many aspects. The human-dominated command, control, and decision-making model of human warfare for thousands of years will be completely transformed. Humans commanding AI, AI commanding humans, and AI commanding AI are all possible scenarios in warfare.

Distributed, networked, flattened, and parallel structures are key characteristics of intelligent combat systems. The centralized, human-centric single-decision-making model is gradually being replaced by decentralized or weakly centralized models based on AI, such as unmanned systems, autonomous swarms, and manned-unmanned collaboration. Hybrid compatibility among these models is becoming a development trend. The lower the operational level and the simpler the mission, the more prominent the role of unmanned and decentralized systems; the higher the level and the more complex the mission, the more important human decision-making and centralized systems become. Pre-war decision-making is primarily human, supplemented by AI; during war, AI is primarily AI, supplemented by human; post-war, both are used, with hybrid decision-making becoming the dominant approach (see Table 3).

In the future battlefield, combat situations will be highly complex, rapidly changing, and exceptionally intense. The convergence of various information sources will generate massive amounts of data, which cannot be processed quickly and accurately by the human brain alone. Only by achieving a collaborative operation mode of “human brain + AI,” based on technologies such as combat cloud, databases, network communication, and the Internet of Things, can “commanders” cope with the ever-changing battlefield and complete command and control tasks. With the increasing autonomy of unmanned systems and the enhancement of swarm and system-wide AI functions, autonomous decision-making is gradually emerging. Once command and control achieve different levels of intelligence, the Out-of-Loop (OODA) loop time will be significantly reduced, and efficiency will be significantly improved. In particular, pattern recognition for network sensor image processing, “optimization” algorithms for combat decision-making, and particle swarm optimization and bee swarm optimization algorithms for autonomous swarms will endow command and control systems with more advanced and comprehensive decision-making capabilities, gradually realizing a combat cycle where “humans are outside the loop.”

Nonlinear amplification and rapid convergence. Future intelligent warfare will no longer be a gradual release of energy and a linear superposition of combat effects, but rather a rapid amplification of multiple effects such as nonlinearity, emergence, self-growth, and self-focusing, and a rapid convergence of results.

Emergence primarily refers to the process by which each individual within a complex system, following local rules and continuously interacting, generates a qualitative change in the overall system through self-organization. In the future, while battlefield information will be complex and ever-changing, intelligent recognition of images, voice, and video, along with processing by military cloud systems, will enable “one-point collection, multi-user sharing.” Through big data technology, it will be rapidly linked with relevant information and integrated with various weapon fire control systems to implement distributed strikes, swarm strikes, and cyber psychological warfare. This will allow for “detection and destruction,” “aggressive attacks at the first sign of trouble,” and “numerical superiority generating psychological panic”—these phenomena constitute the emergence effect.

The emergent effects of intelligent warfare are mainly reflected in three aspects: first, the acceleration of the kill chain caused by the speed of AI decision-making chain; second, the combat effect caused by the numerical advantage of manned and unmanned collaborative systems, especially swarm systems; and third, the rapid swarm emergence behavior based on network interconnection.

As military intelligence develops to a certain stage, the combined effects of advanced AI, quantum computing, IPv6, and hypersonic technologies will result in combat systems exhibiting nonlinear, asymmetric, self-growing, rapid-response, and uncontrollable amplification and operational effects. This is particularly evident in unmanned, swarm, cyber warfare, and cognitive confrontation. The emergence of intelligence from collective ignorance, increased efficiency through sheer numbers, nonlinear amplification, and other emergent effects will become increasingly prominent. AI-driven cognitive, informational, and energy confrontations will intertwine and rapidly converge around a target, with time becoming increasingly compressed and the speed of confrontation accelerating. This will manifest as a dramatic amplification of multiple effects and a rapid convergence of outcomes. Energy shockwaves, rapid-fire combat, AI terminators, public opinion reversals, social unrest, psychological breakdowns, and the chain reaction of the Internet of Things will become prominent characteristics of intelligent warfare.

In unmanned swarm attacks, assuming roughly the same platform performance, the Lanchester equation applies: combat effectiveness is proportional to the square of the number of units; quantity advantage translates to quality advantage. Network attack and defense, and psychological and public opinion effects, follow Metcalfe’s Law, being proportional to the square of the number of interconnected users, with nonlinear and emergent effects becoming more pronounced. The quantity and intelligence of battlefield AI determine the overall level of intelligence in the combat system, impacting battlefield intelligence control and influencing the outcome of war. In the era of intelligent warfare, how to manage the interrelationships between energy, information, cognition, quantity, quality, virtuality, and physicality, and how to skillfully design, control, utilize, and evaluate nonlinear effects, are major new challenges and requirements for future warfare.

In the future, whether it is a reversal of public opinion, psychological panic, swarm attacks, mass operations, or autonomous combat by humans outside the ring, their emergence effects and strike effects will become relatively common phenomena and easy-to-implement actions, forming a capability that is compatible with deterrence and actual combat. It is also a form of warfare that human society must strictly manage and control.

An organically symbiotic relationship between humans and equipment. In the era of intelligence, the relationship between humans and weapons will undergo fundamental changes, becoming increasingly distant physically but increasingly closer in thought. The form of equipment and its development and management models will be completely transformed. Human thought and wisdom will be deeply integrated with weaponry through AI, fully integrated in the early stages of equipment development, optimized and iterated during the use and training phase, and further upgraded and improved after combat verification, in a continuous cycle of progress.

First, with the rapid development of technologies such as network communication, mobile internet, cloud computing, big data, machine learning, and bionics, and their widespread application in the military field, the structure and form of traditional weapons and equipment will be completely changed, exhibiting diverse functions such as front-end and back-end division of labor and cooperation, efficient interaction, and adaptive adjustment. They will be complex entities integrating mechanics, information, networks, data, and cognition.

Secondly, while humans and weapons are gradually becoming physically detached, they are also becoming increasingly integrated into an organic symbiotic entity in terms of mindset. The gradual maturation of drones and robots is shifting their focus from assisting humans in combat to replacing them, with humans taking a more backseat. The integration of humans and weapons will take on entirely new forms. Human thought and wisdom will participate in the entire lifecycle of design, research and development, production, training, use, and support. Unmanned combat systems will perfectly combine human creativity and intellect with the precision, speed, reliability, and fatigue resistance of machines.

Third, profound changes are taking place in equipment development and management models. Mechanized equipment becomes increasingly outdated with use, while information technology software becomes increasingly new, and intelligent algorithms become increasingly sophisticated with use. Traditional mechanized equipment is delivered to the troops using a “pre-research—development—finalization” model, resulting in a decline in combat performance over time and vehicle hours. Information technology equipment is a product of the combined development of mechanization and informatization; the platform remains the same, but the information system is constantly iterated and updated with the development of computer CPUs and storage devices, exhibiting a step-by-step development characteristic of “information-led, software-driven hardware, rapid replacement, and spiral ascent.” Intelligent equipment, based on mechanization and informatization, continuously optimizes and improves training models and algorithms with the accumulation of data and experience, showing an upward curve of becoming stronger and better with use over time and frequency. Therefore, the development, construction, use, training, and support models for intelligent equipment will undergo fundamental changes.

Evolving through learning and confrontation. Evolution will undoubtedly be a defining characteristic of future intelligent warfare and combat systems, and a commanding height in future strategic competition. Combat systems in the intelligent era will gradually acquire adaptive, self-learning, self-confrontational, self-repairing, and self-evolving capabilities, becoming an evolvable ecosystem and game-theoretic system.

The most distinctive and unique feature of intelligent combat systems lies in the combination of human-like and human-like intelligence with the advantages of machines, achieving “superhuman” combat capabilities. The core of this capability is that numerous models and algorithms improve and refine with use, possessing an evolutionary function. If future combat systems resemble the human body, with the brain as the command and control center, the nervous system as the network, and the limbs as weapons and equipment controlled by the brain, like a living organism, possessing self-adaptive, self-learning, self-defense, self-repair, and self-evolutionary capabilities, then we believe it possesses the ability and function of evolution. Because intelligent combat systems are not entirely the same as living organisms, while a single intelligent system is similar to a living organism, a multi-system combat system is more like an “ecosystem + adversarial game system,” more complex than a single living organism, and more adversarial, social, collective, and emergent.

Preliminary analysis suggests that with the development and application of technologies such as combat simulation, virtual reality, digital twins, parallel training, intelligent software, brain-inspired chips, brain-like systems, bionic systems, natural energy harvesting, and novel machine learning, future combat systems can gradually evolve from single-function, partial-system evolution to multi-functional, multi-element, multi-domain, and multi-system evolution. Each system will be able to rapidly formulate response strategies and take action based on changes in the battlefield environment, different threats, different adversaries, and its own strengths and capabilities, drawing upon accumulated experience, extensive simulated adversarial training, and models and algorithms built through reinforcement learning. These strategies will then be continuously revised, optimized, and self-improved through practical warfare. Single-mission systems will possess characteristics and functions similar to living organisms, while multi-mission systems, like species in a forest, will have a cyclical function and evolutionary mechanism of mutual restraint and survival of the fittest, possessing the ability to engage in game-theoretic confrontation and competition under complex environmental conditions, forming an evolvable ecological and game-theoretic system.

The evolution of combat systems mainly manifests in four aspects: First, the evolution of AI. With the accumulation of data and experience, it will inevitably be continuously optimized, upgraded, and improved. This is relatively easy to understand. Second, the evolution of combat platforms and cluster systems, mainly moving from manned control to semi-autonomous and autonomous control. Because it involves not only the evolution of platform and cluster control AI, but also the optimization and improvement of related mechanical and information systems, it is relatively more complex. Third, the evolution of mission systems, such as detection systems, strike systems, defense systems, and support systems. Because it involves multiple platforms and multiple missions, the factors and elements involved in the evolution are much more complex, and some may evolve quickly, while others may evolve slowly. Fourth, the evolution of the combat system itself. Because it involves all elements, multiple missions, cross-domain operations, and confrontations at various levels, its evolutionary process is extremely complex. Whether a combat system can evolve cannot rely entirely on its own growth; it requires the proactive design of certain environments and conditions, and must follow the principles of biomimicry, survival of the fittest, mutual restraint, and full-system lifecycle management to possess the function and capability for continuous evolution.

Intelligent design and manufacturing. In the era of intelligentization, the defense industry will shift from a relatively closed, physical-based, and time-consuming research and manufacturing model to an open-source, intelligent design and manufacturing model that can rapidly meet military needs.

The defense industry is a strategic industry of the nation, a powerful pillar of national security and defense construction. In peacetime, it primarily provides the military with advanced, high-quality, and reasonably priced weaponry and equipment. In wartime, it is a crucial force for operational support and a core pillar for ensuring victory. The defense industry is a high-tech intensive sector. The research and development and manufacturing of modern weaponry and equipment are technology-intensive, knowledge-intensive, systemically complex, and highly integrated. The development of weapons and equipment such as large aircraft carriers, fighter jets, ballistic missiles, satellite systems, and main battle tanks typically takes ten, twenty, or even more years before finalization and delivery to the armed forces, involving large investments, long cycles, and high costs. From the post-World War II period to the end of the last century, the defense industrial system and capability structure were products of the mechanized era and warfare. Its research, testing, manufacturing, and support were primarily geared towards the needs of the military branches and industry systems, mainly including weaponry, shipbuilding, aviation, aerospace, nuclear, and electronics industries, as well as civilian supporting and basic industries. After the Cold War, the US defense industry underwent strategic adjustments and mergers and reorganizations, generally forming a defense industrial structure and layout adapted to the requirements of informationized warfare. The top six defense contractors in the United States can provide specialized combat platforms and systems for relevant branches of the armed forces, as well as overall solutions for joint operations, making them cross-service and cross-domain system integrators. Since the beginning of the 21st century, with the changing demands of system-of-systems and information-based warfare and the development of digital, networked, and intelligent manufacturing technologies, the traditional development model and research and production capabilities of weapons and equipment have begun to gradually change, urgently requiring reshaping and adjustment in accordance with the requirements of informationized warfare, especially intelligent warfare.

In the future, the defense science and technology industry will, in accordance with the requirements of joint operations, all-domain operations, and the integrated development of mechanization, informatization, and intelligence, shift from the traditional focus on service branches and platform construction to cross-service and cross-domain system integration. It will also shift from relatively closed, self-contained, independent, fragmented, physical-based, and long-cycle research, design, and manufacturing to open-source, democratic crowdsourcing, virtual design and integration verification, adaptive manufacturing, and rapid fulfillment of military needs (see Figure 8). This will gradually form a new innovation system and intelligent manufacturing system that combines hardware and software, virtual and real interaction, intelligent human-machine-object-environment interaction, effective vertical industrial chain connection, horizontal distributed collaboration, and military-civilian integration. Joint design and demonstration by multiple military and civilian parties, joint research and development by supply and demand sides for construction and use, iterative optimization based on parallel military systems in both virtual and real environments, and improvement through combat training and real-world verification—a model of simultaneous research, testing, use, and construction—is the basic mode for the development and construction of intelligent combat systems and the generation of combat power.

Wu Mingxi 8

The risk of spiraling out of control. Since intelligent warfare systems theoretically possess the ability to self-evolve and reach “superhuman” levels, if humans do not pre-design control programs, control nodes, and a “stop button,” the result could very well be destruction and disaster. A critical concern is that numerous hackers and malicious warmongers may exploit intelligent technology to design uncontrollable warfare programs and combat methods, allowing numerous machine brains (AIs) and swarms of robots to fight adaptively and self-evolving according to pre-set combat rules, becoming invincible and relentlessly advancing, ultimately leading to an uncontrollable situation and irreparable damage. This is a major challenge facing humanity in the process of intelligent warfare and a crucial issue requiring research and resolution. This problem needs to be recognized and prioritized from the perspective of a shared future for all humanity and the sustainable development of human civilization. It requires designing rules of war, formulating international conventions, and regulating these systems technically, procedurally, ethically, and legally, implementing mandatory constraints, checks, and management.

The above ten transformations and leaps constitute the main content of the new form of intelligent warfare. Of course, the development and maturity of intelligent warfare is not a castle in the air or a tree without roots, but is built upon mechanization and informatization. Without mechanization and informatization, there is no intelligence. Mechanization, informatization, and intelligence form an organic whole, interconnected and mutually reinforcing, iteratively optimizing and leapfrog developing. Currently, mechanization is the foundation, informatization is the guiding principle, and intelligence is the direction. Looking to the future, mechanization will remain the foundation, informatization will provide support, and intelligence will be the guiding principle.

A Bright Future

In the time tunnel of the new century, we see the train of intelligent warfare speeding along. Will humanity’s greed and technological might lead us into a more brutal darkness, or will it propel us towards a more civilized and enlightened future? This is a major philosophical question that humanity needs to ponder. Intelligentization is the future, but it is not everything. Intelligentization can handle diverse military tasks, but it is not omnipotent. Faced with sharp contradictions between civilizations, religions, nations, and social classes, and with extreme events such as thugs wielding knives, suicide bombings, and mass riots, the role of intelligentization remains limited. Without resolving global political imbalances, unequal rights, unfair trade, and social contradictions, war and conflict will be inevitable. Ultimately, the world is determined by strength, and technological, economic, and military strength are extremely important. While military strength cannot determine politics, it can influence it; it cannot determine the economy, but it can bring security for economic development. The stronger the intelligent warfare capabilities, the stronger its deterrent and war-preventing function, and the greater the hope for peace. Like nuclear deterrence, it plays a crucial role in preventing large-scale wars to avoid terrible consequences and uncontrolled disasters.

The level of intelligence in warfare, in a sense, reflects the progress of civilization in warfare. The history of human warfare, initially a struggle between groups for food and habitation, has evolved into land occupation, resource plunder, expansion of political power, and domination of the spiritual world—all fraught with bloodshed, violence, and repression. As the ultimate solution to irreconcilable contradictions in human society, war’s ideal goal is civilization: subjugation without fighting, minimal resource input, minimal casualties, and minimal damage to society… However, past wars have often failed to achieve this due to political struggles, ethnic conflicts, competition for economic interests, and the brutality of technological destructive methods, frequently resulting in the utter destruction of nations, cities, and homes. Past wars have failed to achieve these ideals, but future intelligent warfare, due to technological breakthroughs, increased transparency, and deeper mutual sharing of economic benefits, especially as the confrontation of human forces gradually gives way to confrontation between robots and AI, will see decreasing casualties, material consumption, and collateral damage. This presents a significant possibility of achieving civilization, offering humanity hope. We envision future warfare gradually transitioning from the mutual slaughter of human societies and the immense destruction of the material world to wars between unmanned systems and robots. This will evolve into deterrence and checks and balances limited to combat capabilities and overall strength, AI confrontations in the virtual world, and highly realistic war games… The energy expenditure of human warfare will be limited to a certain scale of unmanned systems, simulated confrontations and experiments, or even merely the energy needed to wage a war game. Humanity will transform from the planners, designers, participants, leaders, and victims of war into rational thinkers, organizers, controllers, observers, and adjudicators. Human bodies will no longer suffer trauma, minds will no longer be frightened, wealth will no longer be destroyed, and homes will no longer be devastated. Although this beautiful ideal and aspiration may always fall short of harsh reality, we sincerely hope that this day will arrive, and arrive as soon as possible. This is the highest stage of intelligent warfare development, the author’s greatest wish, and humanity’s beautiful vision!

(Thanks to my colleague, Researcher Zhou Xumang, for his support and assistance in writing this paper. He has unique thoughts and insights into the development and construction of intelligent systems.)

Notes

[1] Robert O. Walker et al., 20YY: War in the Age of Robots, translated by Zou Hui et al., Beijing: National Defense Industry Press, 2016, p. 148.

The Era of Intelligent War Is Coming Rapidly

Wu Mingxi

Abstract: Since the entry into the new century, the rapid development of intelligent technology with artificial intelligence (AI) at the core has accelerated the process of a new round of military revolution. The competition in the military field is going rapidly to the era of intelligent power. The operational elements represented by “AI, cloud, network, group and end” and their diverse combinations constitute a new battlefield ecosystem, and the winning mechanism of war has changed completely. multiplier, transcendence and active role. The platform has AI control, the cluster has AI guidance, and the system has AI decision-making. The traditional human-based combat method is replaced by AI models and algorithms, and intelligent dominance becomes the core of future war. The stronger the intelligent combat capability, the more hopeful the soldiers may win the war without firing a shot.

現代國語：

2021-08-18 18:53 来源： 《人民论坛·学术前沿》5月下 作者： 吴明曦

【摘要】新世纪以来，以人工智能（AI）为核心的智能科技快速发展，加快了新一轮军事革命的进程，军事领域的竞争正加速走向智权时代。以“AI、云、网、群、端”为代表的作战要素与多样化组合，构成了新的战场生态系统，战争的制胜机理完全改变。基于模型和算法的AI系统将是核心作战能力，贯穿各个方面、各个环节，起到倍增、超越和能动的作用，平台有AI控制，集群有AI引导，体系有AI决策，传统以人为主的战法运用被AI的模型和算法所替代，制智权成为未来战争的核心制权。智能化作战能力越强大，不战而屈人之兵就越有希望。

【关键词】人工智能 无人化 战场生态 战争形态

【中图分类号】TP18 【文献标识码】A

【DOI】10.16619/j.cnki.rmltxsqy.2021.10.005

【作者简介】吴明曦，中国兵器首席科学家、研究员，中国兵器工业集团科技委副秘书长，中国兵器科学研究院科技委副主任。研究方向为国防科技和武器装备发展战略与规划、政策与理论、管理与改革研究。主要著作有《智能化战争——AI军事畅想》等。

智权时代竞争

人类文明的历史，是认识自然、改造自然的历史，也是认识自我、解放自我的历史。人类通过发展科学技术、开发和运用工具，不断增强能力、减轻负担、摆脱束缚、解放自己。战争的控制权也随着科技的进步、人类活动空间的拓展、时代的发展而不断变化、不断丰富和不断演进。19世纪以来，人类先后经历了陆权、海权、空权、天权、信息权的控制与争夺。随着人工智能（AI）、大数据、云计算、生物交叉、无人系统、平行仿真等智能科技的迅速发展及其与传统技术的深度融合，从认识论、方法论和运行机理上，改变了人类认识和改造自然的能力，正在加快推动机器智能、仿生智能、群体智能、人机融合智能和智能感知、智能决策、智能行动、智能保障以及智能设计、研发、试验、制造等群体性重大技术变革，加速战争形态向智权的控制与争夺演变。

智能科技迅速发展，受到世界主要国家的高度重视，成为支撑军事能力跨越发展的强大动力。美俄已将智能科技置于维持其全球军事大国战略地位的核心，其发展理念、发展模式、组织方式、创新应用等已发生重大转变，并开展了军事智能化的实质性应用与实践（见图1）。

2017年8月，美国国防部表示，未来人工智能战争不可避免，美国需要“立即采取行动”加速人工智能战争科技的开发工作。美军提出的“第三次抵消战略”认为，以智能化军队、自主化装备和无人化战争为标志的军事变革风暴正在到来；为此，他们已将自主系统、大数据分析、自动化等为代表的智能科技列为主要发展方向。2018年6月，美国国防部宣布成立联合人工智能中心，该中心在国家人工智能发展战略的牵引下，统筹规划美军智能化军事体系建设。2019年2月，时任美国总统特朗普签署《美国人工智能倡议》行政令，强调美国在人工智能领域保持持续领导地位对于维护美国的经济和国家安全至关重要，要求联邦政府投入所有资源来推动美国人工智能领域创新。2021年3月，美国人工智能国家安全委员会发布研究报告，指出：“自第二次世界大战以来，作为美国经济和军事力量支柱的技术优势首次受到威胁。如果当前的趋势不改变，中国就拥有未来十年内超越美国成为人工智能全球领导者的力量、人才和雄心。”报告认为，美国为维护国家安全和提升国防能力，必须迅速而负责任地使用人工智能，为抵御这些威胁作好准备。报告得出结论，人工智能将改变世界，美国必须发挥带头作用。

俄罗斯也高度重视人工智能的技术发展及其军事运用。俄军方普遍认为，人工智能将引发继火药、核武器之后军事领域的第三次革命。俄罗斯总统普京2017年9月公开提出，人工智能是俄罗斯的未来，谁能成为该领域的领导者，谁就将主宰世界。2019年10月，普京批准《2030年前俄罗斯国家人工智能发展战略》，旨在加快推进俄罗斯人工智能发展与应用，谋求在人工智能领域的世界领先地位。

中国国务院2017年7月印发《新一代人工智能发展规划》，提出了面向2030年新一代人工智能发展的指导思想、战略目标、重点任务和保障措施，部署构筑人工智能发展的先发优势，加快建设创新型国家和世界科技强国。

世界其他主要国家和军事大国，也纷纷推出各自的人工智能发展规划，表明全球范围内围绕“智权”的争夺已经全面展开。陆权、海权、空权、天权、信息权、智权等，都是科技进步的结果、时代的产物，都有各自的优势，也有各自的不足，并且有些理论随着时代的变化，又在不断拓展。从近代以来战争的控制权发展趋势可以看出，信息权与智权是涉及全局的，其权重更重，影响力更大。未来，随着智能化发展步伐的加快，智权将成为一种快速增长的、对作战全局有更大战略影响力的新型战场控制权。

军事智能的本质是利用智能科技为战争体系建立多样化识别、决策和控制模型。这些模型就是人工智能（AI），是新时代智权争夺的核心。其中，战争体系包括：单装、集群、有人无人协同、多域与跨域作战等装备系统；单兵、班组、分队、合成作战单元、战区联指等作战力量；网络化感知、任务规划与指控、力量协同、综合保障等作战环节；网络攻防、电子对抗、舆情控制、基础设施管控等专业系统；智能化设计、研发、生产、动员、保障等军工能力。AI以芯片、算法和软件等形式，嵌入战争体系的各个系统、各个层次、各个环节，是一个体系化的大脑。AI虽然是战争体系的一个局部，但由于其“类脑”功能和“超越人类极限”的能力越来越强，必将主宰未来战争全局。

战场生态重构

传统战争作战要素相对独立、相对分离，战场生态系统比较简单，主要包括人、装备和战法等。智能时代的战争，各作战要素之间融合、关联、交互特征明显，战场生态系统将发生实质性变化，形成由AI脑体系、分布式云、通信网络、协同群、各类虚实端等构成的作战体系、集群系统和人机系统，简称“AI、云、网、群、端”智能化生态系统（见图2）。其中，AI居于主导地位。

AI脑体系。智能化战场的AI脑体系，是一个网络化、分布式的体系，是与作战平台和作战任务相生相伴、如影随形的，其分类方法有多种。按功能和计算能力分，主要包括小脑、群脑、中脑、混合脑和大脑等；按作战任务和环节分，主要包括传感器AI、作战任务规划和决策AI、精确打击和可控毁伤AI、网络攻防AI、电子对抗AI、智能防御AI和综合保障AI等；按形态分，主要包括嵌入式AI、云端AI和平行系统AI等。

小脑，主要指传感器平台、作战平台和保障平台的嵌入式AI，主要执行战场环境探测、目标识别、快速机动、精确打击、可控毁伤、装备保障、维修保障和后勤保障等任务。

群脑，主要指地面、空中、海上、水中和太空无人化集群平台智能控制的AI，主要执行战场环境协同感知、集群机动、集群打击和集群防御等任务，重点包括同构集群系统的算法和有人无人协同等异构系统的算法。

中脑，主要指战场前沿一线分队指挥中心、数据中心、指挥所边缘计算的AI系统，主要执行在线和离线条件下战术分队作战任务动态规划、自主决策与辅助决策。

混合脑，主要指成建制部队作战中，指挥员与机器AI协同指挥和混合决策系统，战前主要执行以人为主的作战任务规划，战中主要执行以机器AI为主的自适应动态任务规划和调整，战后主要执行面向反恐和防卫的混合决策等任务。

大脑，主要指战区指挥中心、数据中心的模型库、算法库、战法库，重点为战役和战略决策起辅助支撑作用。由于数据充足，战场各类AI脑系统，都可以在此进行训练和建模，待成熟时再加载到各个任务系统中。

未来战场，还将有其他不同功能、不同种类、大大小小的AI，如传感器AI，主要包括图像识别、电磁频谱识别、声音识别、语音识别、人类活动行为识别等。随着智能化的快速发展和广泛应用，全社会都会存在大大小小的AI，平时为民众和社会服务，战时完全有可能为军事服务。

分布式云。军事云与民用云有所不同。一般来讲，军事云平台是利用通信网络搜索、采集、汇总、分析、计算、存储、分发作战信息和数据的分布式资源管理系统。军事云平台通过构建分布式系统、多点容错备份机制，具备强大的情报共享能力、数据处理能力、抗打击和自修复能力，可提供固定与机动、公有与私有的云服务，实现“一点采集，大家共享”，大大减少信息流转环节，使指挥流程扁平、快速，避免各级重复分散建设。

从未来智能化战争需求看，军事云至少需要构建战术前端云、部队云、战区云和战略云四级体系。按作战要素也可分为情报云、态势云、火力云、信息作战云、保障云、星云等专业化云系统。

1.前端云，主要是指分队、班组、平台之间的信息感知、目标识别、战场环境分析和行动自主决策与辅助决策，以及作战过程和效果评估等计算服务。前端云的作用主要体现在两个方面。一是平台之间计算、存储资源的相互共享和协同、智能作战信息的互动融合。例如，一旦某一平台或终端被攻击，相关的感知信息、毁伤状况和历史情况，就会通过网络化的云平台自动备份、自动替换、自动更新，并把相关信息上传到上级指挥所。二是离线终端的在线信息服务和智能软件升级。

2.部队云，主要指营、旅一级作战所构建的云系统，重点是针对不同的威胁和环境，开展智能感知、智能决策、自主行动和智能保障等计算服务。部队云建设的目标是要建立网络化、自动备份，并与上级多个链路相连的分布式云系统，满足侦察感知、机动突击、指挥控制、火力打击、后装保障等不同力量的计算需要，满足战术联合行动、有人／无人协同、集群攻防等不同作战任务的计算需要。

3.战区云，重点是提供整个作战区域的战场气象、地理、电磁、人文、社会等环境因素和信息数据，提供作战双方的兵力部署、武器装备配备、运动变化、战损情况等综合情况，提供上级、友军和民用支援力量等相关信息。战区云应具备网络化、定制化、智能化等信息服务功能，并通过天基、空中、地面、海上和水下等军用通信网络，以及采取保密措施下的民用通信网络，与各个作战部队互联互通，确保提供高效、及时、准确的信息服务。

4.战略云，主要是由一个国家国防系统和军队指挥机关建立起来的以军事信息为主，涵盖相关国防科技、国防工业、动员保障、经济和社会支撑能力，以及政治、外交、舆论等综合性的信息数据，提供战争准备、作战规划、作战方案、作战进程、战场态势、战况分析等核心信息及评估分析和建议；提供战略情报、作战对手军事实力和战争动员潜力等支撑数据。

上述各个云之间，既有大小关系、上下关系，也有横向协作、相互支撑、相互服务的关系。军事云平台的核心任务有两个：一是为构建智能化作战的AI脑体系提供数据和计算支撑；二是为各类作战人员和武器平台，提供作战信息、计算和数据保障。此外，从终端和群体作战需求来看，还需要把云计算的一些结果、模型、算法，事先做成智能芯片，嵌入武器平台和群终端，之后，可以在线升级，也可以离线更新。

通信网络。军用通信与网络信息，是一个复杂的超级网络系统。由于军事力量主要是在陆、海、空、天和野战机动、城镇等环境下作战，其通信网络包括战略通信与战术通信、有线通信与无线通信、保密通信和民用通信等。其中，无线、移动、自由空间通信网络是军用网络体系最重要的组成部分，相关的综合电子信息系统也是依托通信网络逐步建立起来的。

机械化时代的军用通信，主要是跟着平台、终端和用户走，专用性得到了满足，但烟囱太多、互联互通能力极差。信息化时代，这种状况开始改变。目前，军用通信网络正在采取新的技术体制和发展模式，主要有两个特征：一是“网数分离”，信息的传输不依赖于某种特定的网络传输方式，“网通即达”，只要网络链路畅通，所需任何信息即可送达；二是互联网化，基于IP地址和路由器、服务器实现“条条大路通北京”，即军用网络化或者栅格化。当然，军事通信网络与民用不同，任何时候都存在战略性、专用性通信需求，如核武器的核按钮通信和战略武器的指挥控制，卫星侦察、遥感和战略预警的信息传输，甚至单兵室内和特种作战等条件下的专用通信，可能仍然采取通信跟着任务走的模式。但即便如此，通用化、互联网化一定是未来军用通信网络发展的趋势，否则不仅造成战场通信频段、电台和信息交流方式越来越多，造成自扰、互扰和电磁兼容困难，无线电频谱管理也越来越复杂，更为重要的是，平台用户之间很难基于IP地址和路由结构等功能来实施自动联通，如同互联网上的电子邮件那样，一键命令可以传给多个用户。未来的作战平台，一定会既是通信的用户终端，也兼有路由器和服务器等功能。

军用通信网络体系主要包括天基通信网、军用移动通信网、数据链、新型通信网、民用通信网等。

1.天基信息网。在天基信息网络建设和天基信息利用方面，美国居于领先地位。因为太空中上千个在轨平台和载荷中，一半多是美国人的。美军在海湾战争后尤其是伊拉克战争期间，通过战争实践加快了天基信息网络的应用和推进步伐。伊拉克战争之后，通过天基信息的利用和基于IP方式互联互通的建立，彻底将海湾战争时期近140个纵向烟囱实现横向互联，大大缩短了“侦察—判断—决策—攻击”（OODA）回路的时间，从天基传感器到射手的时间由海湾战争时的几十个小时缩短到目前采用人工智能识别后仅20秒左右。

随着小卫星技术的飞速发展，低成本、多功能的小卫星越来越多。商用发射随着竞争越来越多，成本也开始急剧下降，并且一次发射可以携带几颗、十几颗甚至几十颗小卫星。如果再将小型化以后的电子侦察、可见光和红外成像，甚至是量子点微型光谱仪都集成在上面，实现侦察、通信、导航和气象、测绘等功能一体化，未来世界和战场将变得更加透明。

2.军用移动通信网。军用移动通信网络主要有三个方面的用途。一是联合作战各军兵种和作战部队之间的指挥控制，这类通信的保密等级较高，可靠性、安全性要求也高。二是平台、集群之间的通信联络，要求具备抗干扰和较高的可靠性。三是武器系统的指控和火控，大多通过数据链解决。

传统的军用移动通信网络，大多是“有中心、纵向为主、树状结构”。随着信息化进程的加快，“无中心、自组网、互联网化”的趋势愈加明显。随着认知无线电技术的逐步成熟和推广（见图3），未来的网络通信系统，能够自动识别战场中的电磁干扰和通信障碍，快速寻找可用频谱资源，通过跳频跳转等方式进行实时通信联络。同时，软件与认知无线电技术还能兼容不同通信频段与波形，便于在旧体制向新体制的过渡中兼容使用。

3.数据链。数据链是一种特殊的通信技术，通过时分、频分、码分等形式，在各作战平台之间实现事先约定的、定期或不定期、有规则或无规则关键信息的传输，只要不被敌方完全掌握或破译，是很难被干扰的。数据链主要分为专用和通用两大类。联合作战、编队协同和集群作战等，主要采用通用数据链。卫星数据链、无人机数据链、弹载数据链、武器火控数据链等，目前多数还是专用的。未来，通用化是一种趋势，专用化将越来越少。此外，从平台和通信的关系来看，平台传感器的信息收发和内部信息处理一般跟着任务系统走，专用化特点较强，平台之间的通信联络和数据传输则越来越通用化。

4.新型通信。传统军用通信以微波通信为主，由于发散角较大，应用平台较多，相应的电子干扰和微波攻击手段发展也较快，容易实施较远距离的干扰与破坏。因此，毫米波、太赫兹、激光通信、自由空间光通信等新型通信手段，就成为既抗干扰，又容易实施高速、大容量、高带宽通信的重要选择。由于高频电磁波发散角较小，虽然抗干扰性能好，但要实现点对点的精确瞄准和全向通信，仍然有一定难度，尤其是在作战平台高速机动和快速变轨条件下，如何实现对准和全向通信，技术上仍在探索之中。

5.民用通信资源。民用通信资源的有效利用，是智能化时代需要重点考虑和无法回避的战略问题。未来通过民用通信网络尤其是5G／6G移动通信，进行开源信息挖掘和数据关联分析，提供战场环境、目标和态势信息，无论是对作战还是非战争军事行动来说都非常重要。在非战争军事行动任务中，尤其是海外维和、救援、反恐、救灾等行动中，军队的专用通信网络，只能在有限范围和地域中使用，而如何与外界交流和联系就成为一个问题。利用民用通信资源，主要有两种途径：一是利用民用卫星特别是小卫星通信资源；二是利用民用移动通信及互联网资源。

军用与民用通信资源的互动利用，核心是要解决安全与保密问题。一种方式是采取防火墙和加密形式，直接利用民用卫星通信和全球移动通信设施来指挥通信和联络，但黑客与网络攻击的风险依然存在。另一种方式是，采用近年发展起来的虚拟化、内联网、半物理隔离、单向传输、拟态防御、区块链等新技术予以解决。

协同群。通过模拟自然界蜂群、蚁群、鸟群及鱼群等行为，研究无人机、智能弹药等集群系统的自主协同机制，完成对敌目标进攻或防御等作战任务，可以起到传统作战手段和方式难以达到的打击效果。协同群是智能化发展的一个必然趋势，也是智能化建设的主要方向和重点领域。单一作战平台，无论战技性能多高、功能多强，也无法形成群体、数量规模上的优势。简单数量的堆积和规模的扩展，如果没有自主、协同、有序的智能元素，也是一盘散沙。

协同群主要包括三个方面：一是依托现有平台智能化改造形成的有人／无人协同群，其中以大、中型作战平台为主构建；二是低成本、同质化、功能单一、种类不同的作战蜂群，其中以小型无人作战平台和弹药为主构建；三是人机融合、兼具生物和机器智能的仿生集群，其中以具有高度自主能力的仿人、仿爬行动物、仿飞禽动物、仿海洋生物为主构建。利用协同群系统实施集群作战特别是蜂群作战，具有多方面的优势与特点。

1.规模优势。庞大的无人系统可以分散作战力量，增加敌方攻击的目标数，迫使敌人消耗更多的武器和弹药。集群的生存能力，因数量足够多而具有较大的弹性和较强的恢复能力，单个平台的生存能力变得无关紧要，而整体的优势更为明显。数量规模使战斗力的衰减不会大起大落，因为消耗一个低成本的无人平台，不像高价值的有人作战平台与复杂武器系统，如B2战略轰炸机，F22、F35先进作战飞机，一旦受到攻击或被击毁，战斗力将急剧下降。集群作战可以同时发起攻击，使敌人的防线不堪重负，因为大部分防御系统能力有限，一次只能处理一定数量的威胁，即便是密集火炮防御，一次齐射也只能击中有限目标，总有漏网之鱼，所以集群系统突防能力极强。

2.成本优势。集群作战特别是蜂群作战大多以中小无人机、无人平台和弹药为主，型谱简单、数量规模较大，质量性能要求相同，便于低成本大规模生产。现代武器装备和作战平台，虽然升级换代的速度明显加快，但成本上涨也极其惊人。二战以后，武器装备研发和采购价格表明，装备成本和价格上涨比性能提升快得多。海湾战争时期的主战坦克是二战时期的40倍，作战飞机和航母则高达500倍。海湾战争之后到2020年，各类主战武器装备价格又分别上涨了几倍、十几倍、甚至几十倍。与此相比，型谱简单的中小无人机、无人平台和弹药具有明显的成本优势。

3.自主优势。在统一的时空基准平台下，通过网络化的主动、被动通信联络和对战场环境目标的智能感知，群体中的单个平台可以准确感知到相互之间的距离、速度和位置关系，也可以快速识别目标威胁的性质、大小、轻重缓急，以及自身与友邻平台距离的远近。在事先制定好作战规则的前提下，可以让一个或数个平台，按照目标威胁的优先级，进行同时攻击和分波次攻击，也可以分组同时攻击、多次攻击（见图4），还可以明确某个平台一旦受损后，后续平台的优先替补顺序，最终达到按照事先约定好的作战规则，自主决策、自主行动。这种智能化作战行动，根据人的参与程度和关键节点控制难度，既可以完全交给群体自主行动，也可以实施有人干预下的半自主行动。

4.决策优势。未来的战场环境日趋复杂，作战双方是在激烈的博弈和对抗中较量。因此，快速变化的环境和威胁，依靠人在高强度对抗环境下参与决策，时间上来不及，决策质量也不可靠。因此，只有交由协同群进行自动环境适应，自动目标和威胁识别，自主决策和协同行动，才能快速地攻击对手或实施有效防卫，取得战场优势和主动权。

协同群给指挥控制带来了新挑战。怎么对集群实施指挥控制是一个新的战略课题。可以分层级、分任务实施控制，大致包括集中控制模式、分级控制模式、一致协同模式、自发协同模式。[1]可以采取多种形式，实现人为的控制和参与。一般来讲，越是在战术层面的小分队行动，越是要采取自主行动和无人干预；在成建制的部队作战层面，由于涉及对多个作战群的控制，需要采取集中规划、分级控制，人要有限参与；在更高级的战略和战役层次，集群只是作为一种平台武器和作战样式来使用，需要统一规划和布局，人为参与的程度就会更高。从任务性质来看，执行战略武器的操作使用，如核反击，就需要由人操作，不适合交给武器系统自主处理；执行重要目标、高价值目标的攻防时，如斩首行动，也需要人全程参与和控制，同时发挥武器系统的自主功能；对于战术目标的进攻，如果需要实施致命打击和毁伤任务的作战行动，可以让人有限参与，或者经人确认后，让协同群去自动执行；执行侦察、监视和目标识别、排查等非打击任务，或执行防空反导等时间短、人难以参与的任务时，主要交由协同群自动执行，而人不需要参与，也无法参与。此外，集群作战也要重视研究其反制措施。重点研究电子欺骗、电磁干扰、网络攻击和高功率微波武器、电磁脉冲炸弹、弹炮系统等反制措施，其相关作用和效果比较明显。同时，还要研究激光武器、蜂群对蜂群等反制措施，逐步建立人类能有效控制的、对付协同群的“防火墙”。

虚实端。虚实端主要指各类与“云、网”链接的终端，包括预先置入智能模块的各类传感器、指控平台、武器平台、保障平台、相关设备设施和作战人员。未来各种装备、平台，都是前台功能多样、后台云端支撑、虚实互动、在线离线结合的赛博实物系统CPS和人机交互系统。在简单环境感知、路径规划、平台机动、武器操作等方面，主要依靠前端智能如仿生智能、机器智能来实现。复杂的战场目标识别、作战任务规划、组网协同打击、作战态势分析、高级人机交互等，需要依靠后端云平台和云上AI提供信息数据与算法支撑。每个装备平台的前端智能与后端云上智能应结合，进行统筹规划与设计，形成前后端一体化智能的综合优势。同时，虚拟士兵、虚拟参谋、虚拟指挥员及其与人类的智能交互、高效互动等，也是未来研究发展的重点与难点。

战争形态质变

近代以来，人类社会主要经历了大规模的机械化战争和较小规模的信息化局部战争。20世纪前半叶发生的两次世界大战，是典型的机械化战争。20世纪90年代以来的海湾战争、科索沃战争、阿富汗战争、伊拉克战争和叙利亚战争，充分体现了信息化战争的形态与特点。新世纪新阶段，随着智能科技的快速发展与广泛应用，以数据和计算、模型和算法为主要特征的智能化战争时代即将到来（见图5）。

机械化是工业时代的产物，技术上以机械动力和电气技术为重点，武器装备形态主要表现为坦克、装甲车辆、大炮、飞机、舰船等，对应的是机械化战争形态。机械化战争，主要基于以牛顿定律为代表的经典物理学和社会化大生产，以大规模集群、线式、接触作战为主，在战术上通常要进行现地侦察、勘查地形、了解对手前沿与纵深部署情况，结合己方能力下定决心，实施进攻或防御，进行任务分工、作战协同和保障，呈现出明显的指控层次化、时空串行化等特点。

信息化是信息时代的产物，技术上以计算机、网络通信等信息技术为重点，装备形态主要表现为雷达、电台、卫星、导弹、计算机、军用软件、指挥控制系统、网电攻防系统、综合电子信息系统等，对应的是信息化战争形态。信息化战争，主要基于计算机与网络三大定律（摩尔定律、吉尔德定律和梅特卡夫定律），以一体化联合、精确、立体作战为主，建立“从传感器到射手的无缝快速信息链接”，夺取制信息权，实现先敌发现与打击。在战术上则要对战场和目标进行详细识别和编目，突出网络化感知和指挥控制系统的作用，对平台的互联互通等信息功能提出了新的要求。由于全球信息系统和多样化网络通信的发展，信息化战争淡化了前后方的界限，强调“侦控打评保”横向一体化和战略、战役、战术的一体化与扁平化。

智能化是知识经济时代的产物，技术上以人工智能、大数据、云计算、认知通信、物联网、生物交叉、混合增强、群体智能、自主导航与协同等智能科技为重点，装备形态主要表现为无人平台、智能弹药、集群系统、智能感知与数据库系统、自适应任务规划与决策系统、作战仿真与平行训练系统、军事云平台与服务系统、舆情预警与引导系统、智能可穿戴系统等，对应的是智能化战争形态。

智能化战争，主要基于仿生、类脑原理和AI的战场生态系统，是以“能量机动和信息互联”为基础、以“网络通信和分布式云”为支撑、以“数据计算和模型算法”为核心、以“认知对抗”为中心，多域融合、跨域攻防，无人为主、集群对抗，虚拟与物理空间一体化交互的全新作战形态。

智能化战争以满足核常威慑、联合作战、全域作战和非战争军事行动等需求为目标，以认知、信息、物理、社会、生物等多域融合作战为重点，呈现出分布式部署、网络化链接、扁平化结构、模块化组合、自适应重构、平行化交互、聚焦式释能、非线性效应等特征，制胜机理颠覆传统，组织形态发生质变，作战效率空前提高，战斗力生成机制发生转变。其实质性的变化主要体现在以下十个方面。

AI主导的制胜机理。在智能化条件下，以“AI、云、网、群、端”为代表的全新作战要素将重构战场生态系统，战争的制胜机理将完全改变。其中，基于模型和算法的AI系统是核心作战能力，贯穿各个方面、各个环节，起到倍增、超越和能动的作用，平台有AI控制，集群有AI引导，体系有AI决策，传统以人为主的战法运用被AI的模型和算法所替代，算法战将在战争中起到决定性作用，作战体系和进程最终将以AI为主导，制智权成为未来战争的核心制权。

不同时代、不同战争形态，战场生态系统是不一样的，作战要素构成、制胜机理完全不同。机械化战争是平台中心战，核心是“动”，主导力量是火力和机动力，追求以物载能、以物释能。作战要素主要包括：人+机械化装备+战法。制胜机理是基于机械化装备作战运用的以人为主导的决策，以多胜少、以大吃小、以快制慢，全面、高效、可持续的动员能力，分别起到决定性或重要的作用。信息化战争是网络中心战，核心是“联”，主导力量是信息力，追求以网聚能、以网释能。作战要素及相互关系主要是：基于网络信息的“人+信息化装备+战法”。信息贯穿于人、装备和战法，建立“从传感器到射手”的无缝信息连接，实现体系化网络化作战能力，以体系对局部、以网络对离散、以快制慢，成为取得战争胜利的重要机理。其中，信息对装备和作战体系起到了倍增的作用，但平台仍然以有人为主，信息围绕人发挥辅助决策的作用，但多数决策还是以人为主。智能化战争是认知中心战，核心是“算”，主导力量是智力，智力所占权重将超过火力、机动力和信息力，追求的将是以智驭能、以智制能，以虚制实、以优胜劣，作战双方谁的AI多，谁的AI更聪明，战场主动权就越大。作战要素及相互关系主要是：AI×（云+网+群+人+装备+战法），可以简化为“AI、云、网、群、端”要素构成的相互关联与融合的战场生态系统。未来，AI在战争中的作用将越来越大、越来越强，最终将发挥决定和主导作用。

强调AI的主导作用，并不否认人在战争中的作用。一方面，人的聪明才智已经前置并赋予了AI；另一方面，在战前、后台和战略层面，在相当长一段时间和可预见的未来，AI是无法取代人类的。

现代战争战场环境越来越复杂、作战对抗速度越来越快，如何快速识别处理海量信息、快速响应战场态势、快速制定决策方案，已远非人力所能，也超出了现有技术手段的极限（见表1、表2）。随着AI在战争体系中的应用越来越广、作用越来越大，作战流程将重新塑造，军事杀伤链将提速增效，感知快、决策快、行动快、保障快，成为未来智能化战争制胜的重要砝码。

未来，通过图像、视频、电磁频谱、语音等智能识别与模式识别，对天空地海传感器网络复杂战场信息能够快速精确实施目标识别。利用大数据技术，通过多源多维定向搜索与智能关联分析，不仅能够对各种打击目标进行准确定位，还能够对人类行为、社会活动、军事行动和舆情态势精准建模，逐步提高预警预测准确率。各战区和战场基于精准战场信息，通过事先虚拟空间的大量平行建模和模拟训练，能够自适应地实施任务规划、自主决策与作战进程控制。各作战平台、集群系统的AI，根据任务规划能够围绕作战目标自主、协同执行任务，并针对随时出现的变化进行能动调整。通过事先建立分布式、网络化、智能化、多模式的保障体系与预置布局，能够快速实施精准物流配送、物资供应和智能维修等。总之，通过智能科技的广泛应用和各种AI系统的能动作用、进化功能，在谋划、预测、感知、决策、实施、控制、保障等作战全过程，实现“简单、快捷、高效、可控”的作战流程再造，能够让人类从繁重的作战事务中逐步解脱出来。作战流程再造将促使未来战场节奏加快、时间压缩、过程变短。

AI主导的制胜机理，主要表现在作战能力、手段、策略和措施方面，全面融合了人的智力，接近了人的智能，超越了人的极限，发挥了机器的优势，体现了先进性、颠覆性和创新性。这种先进与创新，不是以往战争简单的延长线和增长量，而是一种质的变化和跃升，是一种高阶特征。这种高阶特征体现为智能化战争具有传统战争形态所不具备的“类脑”功能和很多方面“超越人类极限的能力”。随着AI的不断优化迭代，它总有一天将超过普通士兵、参谋、指挥员甚至精英和专家群体，成为“超级脑”和“超级脑群”。这是智能化战争的核心和关键，是认识论和方法论领域的技术革命，是人类目前可预见、可实现、可进化的高级作战能力。

虚拟空间作用上升。随着时代的进步和科技的发展，作战空间逐步从物理空间拓展到虚拟空间。虚拟空间在作战体系中的地位作用逐步上升且越来越重要，越来越同物理空间和其他领域实现深度融合与一体化。虚拟空间是由人类构建的基于网络电磁的信息空间，它可以多视角反映人类社会和物质世界，同时可以超越客观世界的诸多限制来利用它。构建它的是信息域，连接它的是物理域，反映出的是社会域，利用它的是认知域。狭义上的虚拟空间主要指民用互联网，广义上的虚拟空间主要指赛博空间（Cyberspace），包括各种物联网、军用网和专用网构成的虚拟空间。赛博空间具有易攻难防、以软搏硬、平战一体、军民难分等特征，已成为实施军事行动、战略威慑和认知对抗的重要战场。

虚拟空间的重要性主要体现在三个方面：一是通过网络信息系统，把分散的作战力量、作战要素连接为一个整体，形成体系化网络化作战能力，成为信息化战争的基础；二是成为网电、情报、舆情、心理、意识等认知对抗的主战场和基本依托；三是建立虚拟战场，开展作战实验，实现虚实互动，形成平行作战和以虚制实能力的核心与关键。

未来，随着全球互联、物联的加速升级，随着天基网络化侦察、通信、导航、移动互联、Wi-Fi和高精度全球时空基准平台、数字地图、行业大数据等系统的建立完善与广泛应用，人类社会和全球军事活动将越来越“透明”，越来越被联网、被感知、被分析、被关联、被控制（见图6），对军队建设和作战呈现全方位、泛在化的深刻影响，智能化时代的作战体系将逐步由封闭向开放、由以军为主向军民融合的“开源泛在”方向拓展。

智能化时代，物理、信息、认知、社会、生物等领域的信息数据将逐渐实现自由流动，作战要素将实现深度互联与物联，各类作战体系将从初级的“能力组合”向高级的“信息融合、数据交链、一体化行为交互”方向发展，具备强大的全域感知、多域融合、跨域作战能力，具备随时随地对重要目标、敏感人群和关键基础设施实施有效控制的能力。美国陆军联合兵种中心的一份报告认为，这个世界正在进入“全球监控无处不在”的时代。即使这个世界无法跟踪所有的活动，技术的扩散也无疑会使潜在的信息来源以指数方式增长。

目前，基于网络的软件攻击已具备物理毁伤能力，军事发达国家的网络攻击已具备入侵、欺骗、干扰、破坏等作战能力，赛博空间已经成为实施军事行动和战略威慑的又一重要战场。美国的网络攻击已经用于实战。突尼斯的本·阿里、利比亚的卡扎菲、伊拉克的萨达姆都曾经被美国的网络攻防和维基解密影响，造成舆情转向、心理失控、社会动荡，导致政权的迅速垮台，对传统战争形态产生了颠覆性影响。通过斯诺登事件，美国使用的11类49项“赛博空间”侦察项目目录清单陆续被曝光，“震网”病毒破坏伊朗核设施、“高斯”病毒群体性入侵中东有关国家、“古巴推特网”控制大众舆情等事件，表明美国已具备对互联网、封闭网络、移动无线网络的强大监控能力、软硬攻击和心理战能力。

战争从虚拟空间实验开始。美军从20世纪80年代就开始了作战仿真、作战实验和模拟训练的探索。后来，美军又率先将虚拟现实、兵棋推演、数字孪生等技术用于虚拟战场和作战实验。据分析，海湾战争、科索沃战争、阿富汗战争、伊拉克战争等军事行动，美军都开展了作战模拟推演，力图找出的最优作战和行动方案。据报道，俄罗斯出兵叙利亚之前，就在战争实验室进行了作战预演，依据实验推演情况，制定了“中央-2015”战略演习计划，针对叙利亚作战演练了“在陌生区域的机动和可到达性”。演习结束后，俄军格拉西莫夫总参谋长强调，以政治、经济及舆论心理战等手段为主，辅之以远程精确的空中打击、特种作战等措施，最终达成政治和战略目的。实践表明，俄出兵叙利亚的进程，与实验、演习基本一致。

未来，随着虚拟仿真、混合现实、大数据、智能软件的应用和发展，通过建立一个平行军事人工系统，使物理空间的实体部队与虚拟空间的虚拟部队相互映射、相互迭代，可以在虚拟空间里解决物理空间难以实现的快速、高强度对抗训练和超量计算，可以与高仿真的“蓝军系统”进行对抗和博弈，不断积累数据，建立模型和算法，从而把最优解决方案用于指导实体部队建设和作战，达到虚实互动、以虚制实、以虚制胜的目的。2019年1月25日，谷歌旗下人工智能团队DeepMind与《星际争霸》开发公司暴雪，公布了2018年12月AlphaSTAR与职业选手TLO、MANA的比赛结果，最终在五局三胜赛制中，AlphaSTAR均以5:0取胜。AlphaSTAR只用了两周时间就完成了人类选手需要200年时间的训练量，展示了在虚拟空间进行仿真对抗训练的巨大优势与光明前景。

无人化为主的作战样式。智能化时代，无人化作战将成为基本形态，人工智能与相关技术的融合发展将逐步把这种形态推向高级阶段。无人系统是人类智慧在作战体系中的充分前置，是智能化、信息化、机械化融合发展的集中体现。无人装备最早出现在无人机领域，1917年，英国造出了世界上第一架无人机，但未用于实战。随着技术发展，无人机逐步用于靶机、侦察、察打一体等领域。进入21世纪以来，无人技术与装备由于具有以任务为中心设计、不必考虑乘员需求、作战效费比高等优势，其探索应用已经实现了巨大跨越，取得了重大突破，显现出快速全方位发展的态势，应用范围迅速拓展，涵盖了空中、水面、水下、地面、空间等各个领域。

近年来，人工智能、仿生智能、人机融合智能、群体智能等技术飞速发展，借助卫星通信与导航、自主导航，无人作战平台能够很好地实现远程控制、编队飞行、集群协同。目前，无人作战飞行器、水下无人平台和太空无人自主操作机器人相继问世，双足、四足、多足和云端智能机器人等正在加速发展，已经步入工程化和实用化快车道，军事应用为期不远。

总体上看，智能化时代的无人化作战，将进入三个发展阶段。第一阶段是有人为主、无人为辅的初级阶段，其主要特点是“有人主导下的无人作战”，也就是事前、事中、事后都是由人完全控制和主导的作战行为。第二阶段是有人为辅、无人为主的中级阶段，其主要特点是“有限控制下的无人作战”，即在作战全过程中人的控制是有限度、辅助性但又是关键性的，多数情况可以依靠平台自主行动能力。第三阶段是规则有人、行动无人的高级阶段，其主要特点是“有人设计、极少控制的无人作战”，人类事先进行总体设计，明确各种作战环境条件下的自主行为与游戏规则，在行动实施阶段主要交由无人平台和无人部队自主执行。

自主行为或者自主性，是无人化作战的本质，是智能化战争既普遍又显著的特征，体现在很多方面。

一是作战平台的自主，主要包括无人机、地面无人平台、精确制导武器、水下和太空机器人等自主能力和智能化水平。

二是探测系统的自主，主要包括自动搜索、跟踪、关联、瞄准和图像、语音、视频、电子信号等信息的智能识别。

三是决策的自主，核心是作战体系中基于AI的自主决策，主要包括战场态势的自动分析、作战任务的自动规划、自动化的指挥控制、人机智能交互等。

四是作战行动的自主协同，前期包括有人无人系统的自主协同，后期包括无人化的自主集群，如各类作战编队集群、蜂群、蚁群、鱼群等作战行为。

五是网络攻防的自主行为，包括各种病毒和网络攻击行为的自动识别、自动溯源、自动防护、自主反击等。

六是认知电子战，自动识别电子干扰的功率、频段、方向等，自动跳频跳转和自主组网，以及面向对手的主动、自动电子干扰等。

七是其他自主行为，包括智能诊断、自动修复、自我保障等。

未来，随着人工智能和相关技术融合发展的不断升级，无人化将向自主、仿生、集群、分布式协同等方向快速发展，逐步把无人化作战推向高级阶段，促使战场上有生力量的直接对抗显著减少。虽然未来有人平台会一直存在，但仿生机器人、类人机器人、蜂群武器、机器人部队、无人化体系作战，在智能化时代将成为常态。由于在众多作战领域都可以用无人系统来替代，都可以通过自主行为去完成，人类在遭到肉体打击和损伤之前，一定有无人化作战体系在前面保驾护航。因此，智能化时代的无人化作战体系，是人类的主要保护屏障，是人类的护身符和挡箭牌。

全域作战与跨域攻防。智能化时代全域作战与跨域攻防，也是一种基本作战样式，体现在很多作战场景、很多方面。从陆、海、空、天到物理、信息、认知、社会、生物多领域，以及虚拟和实体的融合互动，从平时的战略威慑到战时的高对抗、高动态、高响应，时间和空间跨度非常大。既面临物理空间作战和虚拟空间网络攻防、信息对抗、舆情引导、心理战等认知对抗，还面临全球安全治理、区域安全合作、反恐、救援等任务，面临网络、通信、电力、交通、金融、物流等关键基础设施的管控。

2010年以来，以信息化智能化技术成果为支撑，美军提出了作战云、分布式杀伤、多域战、算法战、马赛克战、联合全域作战等概念，目的是以体系对局部、以多能对简能、以多域对单域、以融合对离散、以智能对非智能，维持战场优势和军事优势。美军2016年提出多域战、2020年提出联合全域作战概念，目的是发展跨军种跨领域的联合作战能力，实现单一军种作战背后都有三军的支持，具备全域对多域、对单域的能力优势。

未来，随着人工智能与多学科交叉融合、跨介质攻防关键技术群的突破，在物理、信息、认知、社会、生物等功能域之间，在陆、海、空、天等地理域之间，基于AI与人机混合智能的多域融合与跨域攻防，将成为智能化战争一个鲜明的特征。

智能时代的多域与跨域作战，将从任务规划、物理联合、松散协同为主，向异构融合、数据交链、战术互控、跨域攻防一体化拓展。

一是多域融合。根据多域环境下不同的战场与对手，按照联合行动的要求把不同的作战样式、作战流程和任务规划出来，尽量统一起来，实现信息、火力、防御、保障和指控的统筹与融合，实现战略、战役和战术各层次作战能力的融合，形成一域作战、多域联合快速支援的能力。

二是跨域攻防。在统一的网络信息体系支撑下，通过统一的战场态势，基于统一标准的数据信息交互，彻底打通跨域联合作战侦控打评信息链路，实现在战术和火控层面军种之间协同行动、跨域指挥与互操作、作战要素与能力的无缝衔接。

三是全程关联。把多域融合和跨域攻防作为一个整体，统筹设计、全程关联。战前，开展情报收集与分析，实施舆论战、心理战、宣传战和必要的网电攻击。战中，通过特种作战和跨域行动，实施斩首、要点破袭和精确可控打击（见图7）。战后，防御信息系统网络攻击、消除负面舆论对民众影响、防止基础设施被敌破坏，从多个领域实施战后治理、舆情控制和社会秩序恢复。

四是AI支持。通过作战实验、模拟训练和必要的试验验证、实战检验，不断积累数据、优化模型，建立不同作战样式与对手的AI作战模型和算法，形成一个智能化的脑体系，更好地支撑联合作战、多域作战和跨域攻防。

人与AI混合决策。智能化战场AI脑体系的不断健全、优化、升级和完善，使其将在许多方面超越人类。几千年来，人类战争以人为主的指挥控制和决策模式将彻底改变，人指挥AI、AI指挥人、AI指挥AI等，都有可能在战争中出现。

分布式、网络化、扁平化、平行化是智能化作战体系的重要特征，有中心、以人为主的单一决策模式，逐步被基于AI的无人化、自主集群、有人无人协同等无中心、弱中心模式所改变，相互之间的混合兼容成为发展趋势。作战层级越低、任务越简单，无人化、无中心的作用越突出；层级越高、任务越复杂，人的决策、有中心的作用越重要。战前以人决策为主、以AI决策为辅，战中以AI决策为主、以人决策为辅，战后两者都有、以混合决策为主（见表3）。

未来战场，作战对抗态势高度复杂、瞬息万变、异常激烈，多种信息交汇形成海量数据，仅凭人脑难以快速、准确处理，只有实现“人脑＋AI”的协作运行方式，基于作战云、数据库、网络通信、物联网等技术群，“指挥员”才能应对瞬息万变的战场，完成指挥控制任务。随着无人系统自主能力的增加，集群和体系AI功能的增强，自主决策逐步显现。一旦指挥控制实现不同程度的智能化，侦察—判断—决策—攻击（OODA）回路时间将大大压缩，效率将明显提升。尤其是用于网络传感器图像处理的模式识别、用于作战决策的“寻优”算法、用于自主集群的粒子群算法和蜂群算法等，将赋予指挥控制系统更加高级、完善的决策能力，逐步实现“人在回路外”的作战循环。

非线性放大与快速收敛。未来的智能化作战，不再是能量的逐步释放和作战效果的线性叠加，而是非线性、涌现性、自生长、自聚焦等多种效应的急剧放大和结果的快速收敛。

涌现主要指复杂系统内每个个体都遵从局部规则，不断进行交互后，以自组织方式产生出整体质变效应的过程。未来，战场信息虽然复杂多变，但通过图像、语音、视频等智能识别和军事云系统处理后，具备“一点采集、大家共享”能力，通过大数据技术与相关信息快速关联，并与各类武器火控系统快速交链后，实施分布式打击、集群打击和网络心理战等，能够实现“发现即摧毁”“一有情况群起而攻之”和“数量优势滋生心理恐慌效应”，这些现象就是涌现效应。

智能化作战的涌现效应主要体现在三个方面：一是基于AI决策链的快速而引发的杀伤链的加速；二是有人无人协同特别蜂群系统数量优势所引发的作战效应；三是基于网络互联互通所产生的快速群体涌现行为。

军事智能化发展到一定阶段后，在高级AI、量子计算、IPV6、高超声速等技术共同作用下，作战体系将具备非线性、非对称、自生长、快速对抗、难以控制的放大效应和行动效果，特别在无人、集群、网络舆情、认知对抗等方面尤为明显，群愚生智、以量增效、非线性放大、涌现效应越来越突出，AI主导下的认知、信息、能量对抗相互交织并围绕着目标迅速聚焦，时间越来越被压缩，对抗速度越来越快，即呈现多种效应的急剧放大和结果的快速收敛。能量冲击波、对抗极速战、AI终结者、舆情反转、社会动荡、心理失控、物联网连锁效应等，将成为智能化战争的显著特征。

无人化集群攻击，作战双方在平台性能大致相同的条件下，遵循兰切斯特方程，作战效能与数量的平方成正比，数量优势就是质量优势。网络攻防和心理舆情效应，遵循梅特卡夫定律，与信息互联用户数的平方成正比，非线性、涌现效应更加明显。战场AI数量的多少和智商的高低，更决定着作战体系智能化的整体水平，关系到战场智权的控制，影响战争胜负和结局。智能化时代，如何处理好能量、信息、认知、数量、质量、虚拟、实体之间的相互关系，如何巧妙地设计、把控、运用和评估非线性效应，是未来战争面临的重大新挑战和新要求。

未来，无论是舆情反转、心理恐慌，还是蜂群攻击、集群行动，以及人在环外自主作战，其涌现效应和打击效果，将成为相对普遍的现象和容易实施的行动，形成威慑与实战兼容的能力，也是人类社会必须严加管理和控制的战争行为。

有机共生的人装关系。在智能化时代，人与武器的关系将发生根本性改变，在物理上越来越远、在思维上越来越近。装备形态和发展管理模式将完全改变，人的思想和智慧通过AI与武器装备深度交链，在装备发展阶段充分前置、在使用训练阶段优化迭代、在作战验证之后进一步升级完善，如此循环往复、不断递进。

第一，随着网络通信、移动互联、云计算、大数据、机器学习和仿生等技术的快速发展及其在军事领域的广泛应用，传统武器装备的结构和形态将彻底改变，呈现出前后台分工协作、高效互动、自适应调整等多样化功能，是集机械、信息、网络、数据、认知于一体的复合体。

第二，人与武器逐渐物理脱离，但在思维上逐步深度融合为有机共生体。无人机、机器人的逐步成熟，从辅助人作战转向代替人作战，人更加退居到后台。人与武器的结合方式，将以崭新形态出现。人的思想和智慧将全寿命周期地参与设计、研发、生产、训练、使用和保障过程，无人作战系统将把人的创造性、思想性和机器的精准性、快速性、可靠性、耐疲劳性完美结合起来。

第三，装备建设与管理模式发生深刻变化。机械化装备越用越旧、信息化软件越来越新、智能化算法越用越精。传统的机械化装备采用“预研—研制—定型”的模式交付部队，战技性能随时间和摩托小时呈下降趋势；信息化装备是机械化、信息化复合发展的产物，平台不变，但信息系统随计算机CPU和存储设备的发展不断迭代更新，呈现“信息主导、以软牵硬，快速更替、螺旋上升”的阶梯式发展特点；智能化装备以机械化、信息化为基础，随着数据和经验的积累，不断地优化提升训练模型和算法，呈现随时间和使用频率越用越强、越用越好的上升曲线。因此，智能化装备发展建设及使用训练保障模式，将发生根本性改变。

在学习对抗中进化。进化，一定是未来智能化战争和作战体系的一个鲜明特点，也是未来战略竞争的一个制高点。智能化时代的作战体系将逐步具备自适应、自学习、自对抗、自修复、自演进能力，成为一个可进化的类生态和博弈系统。

智能化作战体系与系统，最大的特点和与众不同之处，就在于其“类人、仿人”的智能与机器优势的结合，实现“超人类”的作战能力。这种能力的核心是众多模型和算法越用越好、越用越精，具备进化的功能。如果未来作战体系像人体一样，大脑是指挥控制中枢，神经系统是网络，四肢是受大脑控制的武器装备，就像一个生命体一样，具备自适应、自学习、自对抗、自修复、自演进能力，我们认为它就具备进化的能力和功能。由于智能化作战体系与生命体不完全一样，单一的智能化系统与生命体类似，但多系统的作战体系，更像一个“生态系统+对抗博弈系统”，比单一的生命体更复杂，更具有对抗性、社会性、群体性和涌现性。

经初步分析判断，随着作战仿真、虚拟现实、数字孪生、平行训练、智能软件、仿脑芯片、类脑系统、仿生系统、自然能源采集和新型机器学习等技术的发展应用，未来的作战体系可以逐步从单一功能、部分系统的进化向多功能、多要素、多领域、多系统的进化发展。各系统能够根据战场环境变化、面临的威胁不同、面临的对手不同、自身具备的实力和能力，按照以往积累的经验知识、大量仿真对抗性训练和增强学习所建立的模型算法，快速形成应对策略并采取行动，进而在战争实践中不断修正、优化和自我完善、自我进化。单一任务系统将具备类似生命体的特征和机能，多任务系统就像森林中的物种群那样具备相生相克、优胜劣汰的循环功能和进化机制，具备复杂环境条件下的博弈对抗和竞争能力，形成可进化的类生态和博弈系统。

作战体系的进化途径，主要体现在四个方面：一是AI的进化，随着数据和经验的积累，一定会不断优化、升级和提升。这一点比较容易理解。二是作战平台和集群系统的进化，主要从有人控制为主向半自主、自主控制迈进。由于不仅涉及平台和集群控制AI的进化，还涉及相关机械与信息系统的优化和完善，所以要相对复杂一点。三是任务系统的进化。如探测系统、打击系统、防御系统、保障系统的进化等，由于涉及多平台、多任务，所以进化涉及的因素和要素就复杂得多，有的可能进化快，有的可能进化慢。四是作战体系的进化，由于涉及全要素、多任务、跨领域，涉及各个层次的对抗，其进化过程就非常复杂。作战体系能否进化，不能完全依靠自生自长，而需要主动设计一些环境和条件，需要遵循仿生原则、适者生存原则、相生相克原则和全系统全寿命管理原则，才能具备持续进化的功能和能力。

智能设计与制造。智能化时代的国防工业，将从相对封闭、实物为主、周期较长的研究制造模式向开源开放、智能设计与制造、快速满足军事需求转变。

国防工业是国家战略性产业，是国家安全和国防建设的强大支柱，平时主要为军队提供性能先进、质量优良、价格合理的武器装备，战时是实施作战保障的重要力量，是确保打赢的核心支撑。国防工业是一个高科技密集的行业，现代武器装备研发和制造，技术密集、知识密集、系统复杂、综合性强，大型航母、战斗机、弹道导弹、卫星系统、主战坦克等武器装备的研发，一般都要经过十年、二十年甚至更长时间，才能定型交付部队，投入大、周期长、成本高。二战以后到上世纪末，国防工业体系和能力结构是机械化时代与战争的产物，其科研、试验、生产制造、保障等，重点面向军兵种需求和行业系统组织科研与生产，主要包括兵器、船舶、航空、航天、核和电子等行业，以及民口配套和基础支撑产业等。冷战后，美国国防工业经过战略调整和兼并重组，总体上形成了与信息化战争体系对抗要求相适应的国防工业结构和布局。美国排名前六位的军工巨头，既可以为相关军兵种提供专业领域的作战平台与系统，也可以为联合作战提供整体解决方案，是跨军兵种跨领域的系统集成商。进入21世纪以来，随着体系化、信息化作战需求的变化和数字化、网络化、智能化制造技术的发展，传统武器装备发展模式和科研生产能力开始逐步改变，迫切需要按照信息化战争特别是智能化战争的要求进行重塑和调整。

未来，国防科技工业将按照联合作战、全域作战、机械化信息化智能化融合发展要求，从传统以军兵种、平台建设为主向跨军兵种、跨领域系统集成转变，从相对封闭、自成体系、各自独立、条块分割、实物为主、周期较长的研究设计制造向开源开放、民主化众筹、虚拟化设计与集成验证、自适应制造、快速满足军事需求转变（见图8），逐步形成软硬结合、虚实互动、人机物环智能交互、纵向产业链有效衔接、横向分布式协同、军民一体化融合的新型创新体系和智能制造体系。军地多方联合论证设计，建设和使用供需双方共同研发，基于平行军事系统的虚实迭代优化，通过作战训练和实战验证来完善提升，边研边试边用边建，是智能化作战体系发展建设和战斗力生成的基本模式。

吴明曦8

失控的风险。由于智能化作战体系在理论上具备自我进化并达到“超人类”的能力，如果人类不事先设计好控制程序、控制节点，不事先设计好“终止按钮”，结果很可能会带来毁灭和灾难。需要高度关注的是，众多黑客和“居心不良”的战争狂人，会利用智能化技术来设计难以控制的战争程序和作战方式，让众多机器脑AI和成群结队的机器人，按照事先设定的作战规则，自适应和自演进地进行战斗，所向披靡，勇往直前，最终酿成难以控制的局面，造成难以恢复的残局。这是人类在智能化战争进程中面临的重大挑战，也是需要研究解决的重大课题。需要从全人类命运共同体和人类文明可持续发展的高度，认识和重视这个问题，设计战争规则，制定国际公约，从技术上、程序上、道德上和法律上进行规范，实施强制性的约束、检查和管理。

以上十个方面的突变和跨越，是智能化战争新形态的主要内容。当然，智能化战争的发展与成熟，并不是空中楼阁、无本之木，而是建立在机械化和信息化之上。没有机械化和信息化，就没有智能化。机械化、信息化、智能化“三化”是一个有机整体，相互联系、相互促进，迭代优化、跨越发展。从目前看，机械化是基础，信息化是主导，智能化是方向。从未来看，机械化是基础，信息化是支撑，智能化是主导。

未来美好远景

在新世纪的时空隧道里，我们看到智能化战争的列车正快速行驶，是任由人类的贪婪和科技的强大走向更加残酷的黑暗，还是迈向更加文明和光明的彼岸，这是人类需要思索的重大哲学命题。智能化是未来，但不是全部。智能化能胜任多样化军事任务，但不是全能。面对文明之间、宗教之间、国家之间、阶层之间的尖锐矛盾，面对手持菜刀的暴徒、自杀式爆炸、群体性骚乱等极端事件，智能化作用仍然有限。全球政治不平衡、权利不平等、贸易不公平、社会矛盾不解决，战争和冲突将不可避免。世界最终靠实力说了算，而其中科技实力、经济实力和军事实力极其重要。军事实力虽然决定不了政治，但可以影响政治，决定不了经济，但可以为经济发展带来安全。智能化作战能力越强大，其威慑强敌、遏制战争的功能越强，和平就越有希望。就像核威慑那样，为避免可怕的后果和失控的灾难，在防止大规模战争方面发挥着重要的作用。

战争的智能化程度，在某种意义上体现了战争文明的进程。人类战争的历史，最初由族群之间食物和居住区域的争夺，到土地占领、资源掠夺、政治实力扩张、精神世界统治，无不充满血腥、暴力和镇压。战争作为人类社会不可调和矛盾的最终解决手段，其所追求的理想目标是文明化：不战而屈人之兵、资源投入最少、人员伤亡最小、对社会的破坏最轻……但以往的战争实践，往往因政治斗争、民族矛盾、经济利益争夺、科技毁伤手段的残酷等原因而事与愿违，常常把国家、城市和家园毁坏殆尽。以往的战争未能实现上述理想，而未来智能化战争由于技术上的突破、透明度的增加、经济利益互利共享的加深，特别是有生力量的对抗逐步让位于机器人之间的对抗、AI之间的博弈，人员伤亡、物质消耗、附带损伤会越来越小，在很大程度上存在实现文明化的可能性，给人类带来了希望。我们期待，未来战争，从人类社会的相互残杀、物质世界的极大破坏，逐步过渡到无人系统和机器人之间的战争，发展到仅限于作战能力和综合实力的威慑与制衡、虚拟世界中AI之间的对抗、高仿真的战争游戏……人类战争的消耗，只限于一定规模的无人系统、模拟对抗与仿真实验，甚至仅仅是打一场战争游戏的能源。人类由战争的谋划者、设计者、参与者、主导者和受害者，转变为理性的思想者、组织者、控制者、旁观者和裁决者。人类的身体不再受到创伤，精神不再受到惊吓，财富不再遭到破坏，家园不再遭到摧毁。虽然美好的理想和愿望，与残酷的现实可能始终存在差距，但衷心希望这一天能够到来，尽早到来。这是智能化战争发展的最高阶段，作者的最大愿望，人类的美好远景！

（感谢同事周旭芒研究员为论文撰写提供支持和帮助，他在智能化发展和建设方面有独到的思想和见解）

注释

[1]［美］罗伯特·O.沃克等：《20YY：机器人时代的战争》，邹辉等译，北京：国防工业出版社，2016年，第148页。

The Era of Intelligent War Is Coming Rapidly

Wu Mingxi

Abstract: Since the entry into the new century, the rapid development of intelligent technology with artificial intelligence (AI) at the core has accelerated the process of a new round of military revolution. The competition in the military field is going rapidly to the era of intelligent power. The operational elements represented by “AI, cloud, network, group and end” and their diverse combinations constitute a new battlefield ecosystem, and the winning mechanism of war has changed completely. The AI system based on models and algorithms will be the core combat capability, running through all aspects and links and playing a multiplier, transcendence and active role. The platform has AI control, the cluster has AI guidance, and the system has AI decision-making. The traditional human-based combat method is replaced by AI models and algorithms, and intelligent dominance becomes the core of future war. The stronger the intelligent combat capability, the more hopeful the soldiers may win the war without firing a shot.

中國原創軍事資源：https://www.rmlt.com.cn/2021/0818/622318889.shtml