人工智慧融入中國軍隊：解放軍戰略支援部隊發展人工智慧創新能力

現代英語：

Artificial Intelligence ( AI ) technology has advanced by leaps and bounds in recent years . All major powers have developed advanced AI capabilities and attempted to effectively integrate AI into their armed forces. Beijing has also released an ambitious plan to make China a global leader in advanced artificial intelligence by 2030. Chinese Communist Party leader Xi Jinping also reiterated at the 20th Party Congress that China should attach equal importance to the development of artificial intelligence and “intelligent warfare.”

Although China’s strategic goals in the field of artificial intelligence are clear, how it will integrate artificial intelligence into the People’s Liberation Army remains opaque. But at least the recently established PLA Strategic Support Force (SSF) provides some clues: the organization has been given an innovative mission and is responsible for integrating multiple “strategic functions.” To effectively understand the Strategic Support Force, we need to explore whether it will have a “game-changing” impact in future conflicts, where mastery of the information domain and effective integration of artificial intelligence may determine victory or defeat.

The PLA’s “Joint Operationality”: Strategic Support Force

The PLA underwent major reforms in 2015, partly motivated by the need to shift the PLA’s force focus from land territorial defense to extended force projection to ensure China’s strategic interests in areas such as space, cyber warfare and the far seas. A key element of these reforms is the creation of the Strategic Support Force, which concentrates tasks in these broad areas.

The Strategic Support Force (SSF) is tasked with integrating many “strategic” functions and capabilities previously dispersed across the PLA, including space, cyber, information, and psychological warfare. Today, the Strategic Support Force consists of two departments covering these functions: the Space Systems Department, which is responsible for all space-related missions; and the Cyber ​​Systems Department, which undertakes the PLA’s broad information warfare activities.

The ultimate goal of the Strategic Support Force is to gain information advantage, achieve decision-making advantage, and thus achieve ultimate victory. Analysts recently concluded that its mission is likely to support the pursuit of information superiority and can be divided into two categories: providing strategic information superiority and support capabilities to the PLA’s top leadership, including counter-space operations and offensive cyber warfare, and providing information support services to theater military commands.

The Strategic Support Force appears to be designed to enhance the PLA’s “jointness,” or its ability to conduct joint operations. Historically, the PLA has faced challenges integrating joint operations due to the difficulty in innovating and implementing new command and control and intelligence, surveillance, and reconnaissance capabilities. Because the information support provided by the Strategic Support Force is likely to include “intelligence, surveillance and reconnaissance to achieve operational and strategic objectives”, some experts believe that the Strategic Support Force plays a key role in improving the overall joint combat effectiveness of the PLA.

The SSF also appears to have several mechanisms in place to develop or acquire technology in order to carry out its mission effectively. While the SSF is not the only agency within the PLA with this function, the SSF is responsible for more AI-related equipment contracts than any other service in the PLA. For example, the Strategic Support Force has made significant investments in artificial intelligence innovation, leveraging citizen partnerships to acquire new technologies such as intelligence, surveillance and reconnaissance, autonomous vehicles, information and electronic warfare, simulation and training, and target identification.

PLA Artificial Intelligence Innovation

The U.S. National Security Council pointed out in its 2022 final report that “while artificial intelligence will be widely used in all fields, the large amount of data associated with space, cyber, and information operations makes these application cases particularly suitable for priority integration of AI technology in war simulations, exercises, and experiments.” This is exactly the area where the Strategic Support Force operates.

Many of the Strategic Support Force’s functions involve processing a variety of diverse, large volumes of rapidly changing information flows at speeds exceeding human capacity, making them excellent candidates for the application of artificial intelligence. For example, AI can help create and maintain situational awareness and can be used for prediction by collecting, integrating and analyzing information. AI can also be used to analyze the consequences and planning of potential actions and conduct war simulations.

However, actual command decisions are made by the theater military command or the Joint Chiefs of Staff. At the same time, the development of AI for decision-support applications may be the responsibility of other PLA components, such as the National University of Defense Technology and the Academy of Military Sciences, rather than the Strategic Support Force. Nevertheless, in providing information support to these decision makers, the Strategic Support Force is likely to play an important role in human-machine interface interaction with such artificial intelligence systems, thereby effectively supporting decision making.

At the same time, of the twelve major military applications currently being developed by the PLA, at least five are closely related to the missions of the Strategic Support Force, namely smart satellites, intelligence, surveillance and reconnaissance software, automated cyber attack software, cognitive electronic software, and possible automated vehicles, including:

1. Space Battle
2. Cyber ​​Warfare
3. Electronic warfare

While many applications of AI within the SSU mission area can be identified today, the most important long-term impacts may be difficult to predict. Furthermore, the AI ​​plans proposed in China’s national white paper are not consistent with actual innovation progress.

In fact, most of China’s major investments in AI appear to be business-related and have little to do with military missions. Some assessments suggest that previous estimates of China’s current AI capabilities may be overstated. This means that it is not possible to immediately see the effective integration of artificial intelligence into the PLA’s mission areas, but the Strategic Support Force does have the mission of gaining information advantages to achieve decision-making advantages and ultimate victory.

As for artificial intelligence, it means that the Strategic Support Force must integrate artificial intelligence applications to make up for the PLA’s weaknesses in ensuring and utilizing information advantages. But this does not mean that the Strategic Support Force will become the focus of the PLA’s overall artificial intelligence innovation.

While many of the SSF’s missions are amenable to AI and there may be synergies between missions for applying AI, it is unclear which applications the SSF will use, whether these synergies are feasible, and whether the SSF has the capability to execute them. There will inevitably be a certain degree of prioritization in innovation for specific AI applications that are differentiated across different tasks.

China is determined to become a global leader in artificial intelligence and apply its technology to military missions to suppress U.S. advantages in the Indo-Pacific region. In many ways, the SSF has the advantages to achieve these goals, including a relaxed policy environment that promotes innovation, the SSF’s clear innovation responsibilities, and senior leadership support for “smartness.” The SSF also builds partnerships with China’s high-tech commercial sector and academia. These efforts are consistent with China’s military-civil fusion agenda, which aims to overcome barriers that prevent the People’s Liberation Army from acquiring resources from the commercial sector.

Obstacles to the Strategic Support Force’s Implementation of the PLA’s Innovation Plan

However, the Strategic Support Force also faces huge obstacles in implementing the People’s Liberation Army’s innovation-driven plan. The SSF and the PLA as a whole will face several challenges in AI applications, including attracting and retaining high-quality high-tech talent and mainland China’s inability to domestically develop and manufacture advanced logic and memory chips that are critical to developing cutting-edge AI—a clear weakness now that the United States has disrupted its supply of high-end semiconductors. In addition, research institutions in the United States and other Western countries are now increasingly cautious about collaborating with Chinese researchers in fields such as artificial intelligence, which have significant military potential.

The PLA’s limited combat experience has led to a lack of relevant “real and empirical” data, which may hinder the development of decision-making support artificial intelligence systems. More importantly, unless the PLA focuses on understandable, trustworthy AI, the use of AI systems with opaque operations, uncertain effective areas, and uncertain failure modes could cause serious damage.

For the national army , although there may be some limitations and uncertainties in the military application of artificial intelligence, with the continuous development and maturity of artificial intelligence technology, its application potential in the military field is still huge. As technology advances, we can expect to see more artificial intelligence systems introduced into military applications to improve operational efficiency and combat effectiveness.

However, to ensure that artificial intelligence technology can be robustly applied to military missions, it is necessary to strengthen technology research and development and testing, ensure the safety and reliability of the system, and rationally plan and manage the use of artificial intelligence technology. Only in this way can we better utilize artificial intelligence technology to enhance the information-based combat effectiveness of our military and achieve stronger and more robust combat capabilities.

現代國語：

近年來人工智能（Artificial Intelligence, AI）技術突飛猛進，所有主要大國都發展先進的AI能力，並試圖將AI有效地融入武裝力量。北京亦發佈雄心勃勃之計畫，要求2030年將中國建設成為全球先進人工智能強國。中共領導人習近平亦於第20次黨代表大會再次強調，中國應於人工智慧領域發展與「智慧戰爭」 並重。

雖然中國大陸於人工智能領域的戰略目標明確，但其如何將人工智能融入解放軍仍然是不透明的。但至少，最近成立的解放軍戰略支援部隊（Strategic Support Force, SSF）提供了一些線索：該組織賦予了創新任務，負責整合多種「戰略功能」。為了有效理解戰略支援部隊，探究它是否將在未來衝突中產生「改變遊戲規則」的影響，其中掌握資訊領域和有效整合人工智能可能決定勝負。

解放軍的「聯合作戰性」：戰略支援部隊

2015年解放軍軍隊進行重大改革，部分動機是將解放軍武力重點從陸地領土防禦轉向延伸武力投射，以確保中國在太空、網絡戰和遠海等領域的戰略利益。這些改革的主要關鍵要素即為成立戰略支援部隊，集中了這些廣泛領域中的任務。

戰略支援部隊（SSF）的任務是整合先前分散在解放軍各部隊的許多「戰略」職能和能力，包括太空、網絡、資訊與心理戰等。如今，戰略支援部隊由兩個部門組成，涵蓋了這些職能：太空系統部門，負責所有與太空有關的任務；網路系統部門，承擔解放軍廣義的資訊戰活動。

戰略支援部隊最終目標是獲得資訊優勢，實現決策優勢，從而取得最終勝利。分析家最近得出結論，其任務很可能是支援追求資訊優勢，並可分為兩類：為解放軍最高層領導提供戰略資訊優勢和支援能力，包括反太空行動和攻擊性網絡戰，以及向戰區軍事指揮部提供資訊支援服務。

戰略支援部隊設計似乎旨在提升解放軍的「聯合作戰性」，或其進行聯合作戰的能力。歷史上，因新的指揮掌控與情報、監視和偵察能力的創新和實施方面實存在困難性，解放軍在整合聯合作戰方面面臨挑戰。因戰略支援部隊提供的資訊支援很可能包含「情報、監視和偵察，以實現作戰和戰略目標」，一些專家認為戰略支援部隊在提高解放軍整體聯合作戰性方面起到了關鍵作用。

為了有效遂行任務，戰略支援部隊似乎還擁有若干機制來開發或取得技術。雖然戰略支援部隊並非解放軍內唯一承擔此功能的機構，戰略支援部隊負責與人工智能相關的裝備合約比解放軍的任何一個服務都要多。例如戰略支援部隊在人工智慧創新方面進行了大量投資，利用公民合作夥伴關係來獲取情報、監視和偵察、自主車輛、資訊與電子戰、模擬與訓練、以及目標識別等新技術。

2015年12月31日，中國人民解放軍陸軍領導機構、中國人民解放軍火箭軍、中國人民解放軍戰略支援部隊成立大會在北京八一大樓隆重舉行。習近平授予陸軍、火箭軍、戰略支援部隊軍旗並致訓詞。 圖／新華社
2015年12月31日，中國人民解放軍陸軍領導機構、中國人民解放軍火箭軍、中國人民解放軍戰略支援部隊成立大會在北京八一大樓隆重舉行。習近平授予陸軍、火箭軍、戰略支援部隊軍旗並致訓詞。 圖／新華社

解放軍人工智慧創新

美國國家安全委員會於2022年最終報告中指出，「雖然人工智能將在所有領域普遍應用，但與空間、網絡和信息作戰領域相關的大數據量，使得這些領域的應用案例特別適合優先整合AI技術應用於戰爭模擬、演習和實驗。」而這正是戰略支援部隊所運營的領域。

戰略支援部隊許多功能涉及處理各種多樣、大量和快速變化之信息流，處理速度超過人類容量，都是應用人工智能的絕佳候選領域。例如，人工智慧可以幫助創建和維持態勢感知，透過收集、整合和分析信息，也可用於預測；另人工智能還可於分析潛在行動的後果與規劃、進行戰爭模擬。

然而實際指揮決策則由戰區軍事指揮部或聯合參謀部負責。同時，用於決策支援應用的人工智慧的開發可能由其他解放軍組成部分，如中國國防科技大學和軍事科學院，並非由戰略支援部隊負責。儘管如此，在為這些決策者提供資訊支援方面，戰略支援部隊很可能在與此類人工智慧係統進行人機界面互動發揮重要作用，並進而有效支援決策制定。

與此同時，解放軍刻正開發的十二個主要軍事應用中，至少五個與戰略支援部隊的任務密切相關，分別是智慧衛星、情報、監視和偵察軟件、自動化網絡攻擊軟件、認知電子軟件以及可能的自動化車輛，包含：

太空戰
對於太空戰，解放軍可能探索應用人工智能，內含用於管理大規模衛星；與在地球觀測資料集分析（處理和目標識別衛星圖像）；認知無線電技術，一種「智慧」無線電技術，旨在藉由自主切換頻道來避免干擾和擁擠，從而使空間通訊更加高效；自主衛星操作，以彌補有限的帶寬和延遲與減輕地面衛星操作員的工作負荷。

網路戰
網路戰，即為網路安全和電子戰都依賴於即時處理大量資料以識別威脅並更新防禦措施。如藉由人工智能引導對敵對網絡的指引，亦可同時支援技術偵察和網絡攻擊。此任務處於戰略支援部隊的職責範圍內，此部隊的創立旨在鞏固解放軍在網路安全和電子戰方面的聯繫。自動化還可以提高網絡防禦的速度和規模，減輕先前解放軍面臨的限制。

電子戰
人工智能亦可協助電子戰，增強電子戰系統的自主性。如人工智能可以識別和分類雷達或通訊系統的信號，檢測和分類幹擾器，提高幹擾信號的到達方向估計，開發高效的抗干擾協議與即時更新防禦措施（認知電子戰）。

示意圖。圖為2018年中國北京舉行的「安全中國」（Security China）上展示的人臉辨識技術。 圖／美聯社

雖然目前可確定戰略支援單位任務範圍內許多人工智能的應用，但最重要的長期影響可能難以預測。此外，中共國家白皮書中提出的人工智慧計劃並非與實際的創新進展一致。

實際上，中共於人工智能大部分重要投資似乎與商業相關，與軍事任務關係不大。一些評估先前關於中國當前人工智能能力的估計可能過於誇大。意謂不可能立即看到人工智能有效整合於解放軍的任務領域，但戰略支援部隊確實賦予提取信息優勢以實現決策優勢和最終勝利的任務。

至於人工智能方面，意謂戰略支援部隊必然整合人工智能應用，以彌補解放軍在確保和利用資訊優勢方面之弱點。但這不意謂戰略支援部隊將成為解放軍整體人工智慧創新的焦點。

儘管戰略支援部隊的許多任務均適用於人工智能，而且在任務之間可能存在應用人工智能的協同效應，但目前尚不清楚戰略支援部隊將使用哪些應用，這些協同效應是否具有可行性，與戰略支援部隊是否有能力執行。對於在不同任務中具有差異的特定人工智慧應用，創新中必然會存在一定程度的優先考慮。

中國決心成為全球人工智慧領域領導者，並將其技術應用於軍事任務，以壓制美國在印太地區的優勢。在許多方面，戰略支援部隊都具備説明實現這些目標的優勢，包括寬鬆的政策環境促進創新、戰略支援部隊明確的創新職責以及高層領導對「智慧化」的支持。戰略支援部隊也與中國的高科技商業部門和學術界建立合作關係。這些努力與中國的軍民融合協議一致，旨在克服妨礙解決中國人民解放軍從商業部門獲取資源屏障。

戰略支援部隊執行解放軍創新計畫的障礙

不過戰略支援部隊在執行中國人民解放軍創新驅動計畫時，亦面臨巨大障礙。戰略支援部隊和解放軍總體上將在人工智能應用面臨一些挑戰，包括吸引和留住高素質高科技人才，與中國大陸無法於國內研發和製造對開發尖端人工智能至關重要的先進邏輯和存儲芯片——這是目前美國已中斷對高端半導體的供應所導致的明顯弱點。此外，美國和其他西方國家的研究機構現在對與具有重大軍事潛力的人工智能等領域的中國研究人員進行合作越來越謹慎。

根據解放軍有限的戰鬥經驗，導致了相關「實踐實證」資料匱乏，可能妨礙發展決策支援人工智慧系統。更重要的是，除非解放軍專注於可理解、值得信賴的人工智能，否則使用具有不透明操作、不確定有效區域和不確定故障模式的人工智能係統可能會造成嚴重損害。

對於國軍而言，儘管人工智能於軍事應用可能存在一些限制和不確定性，但隨著人工智能技術的不斷發展和成熟，其在軍事領域的應用潛力仍然巨大。隨著技術的進步，我們有望看到更多人工智慧係統被引入軍用應用中，提高作戰效率和戰鬥力。

然而，要確保人工智能技術能夠穩健地應用於軍事任務中，需要加強技術研發和測試，確保系統的安全性和可靠性，以及合理規劃和管理人工智能技術的使用。只有這樣，我們才能更好地利用人工智慧技術來提升國軍資訊化戰鬥力，並實現更強大、更穩健的作戰能力。

中國原創軍事資源：https://opinion.udn.com/opinion/story/123525/7360758884