AI For Unified Command Plan

In the Future . . .

1. Global competition will force the United States to adopt new combatant command designs optimized for gray zone competition and security cooperation. These new commands will better integrate gender dynamics and analyze human terrain and competing influence networks to gain access, deny benefits, impose costs, and reassure partners.

2. The integrated deterrence doctrine will drive the U.S. military to combine cyber, space, and electronic warfare and other information-related capabilities into a new multidomain strategic command. The command will support integrated planning, provide a larger menu of options for conventional and nuclear deterrence, and synchronize multidomain effects.

3. There will be fewer combatant commands but more combined joint interagency task forces scaled to meet emerging challenges. The United States will check authoritarian advances globally by better synchronizing instruments of power and creating new command and control architectures to align authorities and partners against missions globally.

Introduction

Every two years, a ritual takes over the Pentagon. A mix of service priorities, presidential prerogatives, defense ideas, and congressional intrigue combine in a cauldron called the Unified Command Plan (UCP). The classified document specifies functional and geographic responsibilities and uses this division to assign missions, planning, training, and operational responsibilities.

As the ritual takes place in the “Puzzle Palace” today, the authors have decided to join the ranks of pundits new and old with ideas for revising the UCP — but with a twist. They have refined a large language model (LLM) with over 600 authoritative texts on topics ranging from military history and international relations theory to competitive strategy and deterrence into nine datasets to explore alternative blueprints for the U.S. military. Two insights have emerged.

First, the tendency to create new commands may have reached its limit and potentially undermines calls for integrated deterrence and better synchronization of effects across domains, the U.S. government, and the U.S. network of partners and allies. For example, in this study, the LLM — across multiple iterations — kept generating UCP designs optimized for long-term competition with China that merged cyber, space, and electronic warfare capabilities into a unified command.

Second, the current U.S. blueprint for projecting power and influence struggles in day-to-day competition and integrating development assistance, security cooperation, and economic measures in a holistic manner. The LLM consistently generated alternative UCP designs organized around this central idea.

The net result is a series of options the authors hope will help the civil servants and military professionals who have the Sisyphean task of aligning the structure of the Department of Defense (DOD) with the reality of twenty-first-century great power competition.

What’s Wrong with the UCP?

Critiques of the UCP focus either on specific issues and geographic seams or on larger bureaucratic challenges associated with the design of the national security enterprise. In fact, War on the Rocks has been a hotbed for debates about the right types of commands to address the changing character of war and evolving strategic challenges. Authors have debated U.S. Space Command, building a future warfare command, and geographic designs that are better aligned with the prevailing demand signal.

A risk-averse culture and bureaucracy haunt these debates. Combatant commands appear as bloated proconsuls because they try to compensate for a misaligned national security enterprise. Organizationally, the U.S. government struggles to reconcile maps between the Department of State (DOS) and DOD because the perception of the world is shaped by theater and global campaign plans. Out of desperation, many stakeholders therefore turn to the National Security Council (NSC) to synchronize interagency activities. However, statutorily speaking, the NSC is a coordinating body to untangle the differences across the federal government rather than policymaking to align campaign plans. The net result is strategy as cacophony and an inherent dilemma where unity of effort should reside.

Another issue is proponency, a fancy term that describes the advocacy of people sitting at senior decisionmaking tables and, crucially, their priorities. Proponency is a key aspect of defense politics that plays out in UCP revisions, among other places. The higher the rank of the person in charge of a particular issue set, region, or capability, the easier it will be to argue for budget shares. In 2019, concerns about insufficient advocacy for the space mission within the DOD led to a proponency overcompensation in the form of the near-simultaneous creation of the Space Force and U.S. Space Command. In 1986, the Goldwater-Nichols Department of Defense Reorganization Act ensured that operational commanders and service chiefs had a clear division of roles and responsibilities in order to prevent further U.S. military fiascos. Needless to say, focusing a service chief and a four-star commander on the same mission set runs counter to the organizational design of the Pentagon and could create greater risk of future military fiascos.

The other problem with proponency dominating the command structure is that there are only so many senior decisionmaking tables, but the list of issues the department grapples with is enormous and grows every day. There is an inherent mismatch between what is important and who gets to decide what to do about things that are deemed important — not to mention the fact that every four-star leader is supported by a galaxy of subordinate general and flag officers. The issue has led at least one former secretary of defense to complain of “brass creep,” or an inflated number of officers with stars on their shoulders running the U.S. military without solving enduring strategic challenges.

In addition to these concerns, like most bureaucratic processes, the result tends to create winners and losers while leaving both equally unhappy. With a two-year update cycle prone to parochial infighting over budgets and authorities, change tends to take the form of addition over optimization. It is easier to satisfice and add structure than it is to realign or reduce the number of combatant commands. This puts a premium on growth that may be suboptimal and reinforces a tendency for service-level interests over a unified joint or combined approach to modern operational art and campaigning. Furthermore, competing ideas, parochial interests, veto players, and the enduring quest for turf and money plague open dialogue. The net result is diminishing returns and a span of control issues that inflate the number of four-star generals and staffs running the U.S. military without solving enduring strategic challenges. As staffs become bloated, command structures grow unwieldy, and defense budgets balloon, roles and missions expand without a corresponding increase in either security or strategic advantage.

To Answer a Strategy Question, Tailor a Model

To analyze options for UCP reform, the authors worked with the CSIS Futures Lab and Scale AI to build datasets optimized for retrieval augmented generation — a technique for refining how large models weight documents to generate text. This technique is particularly useful for overcoming the inherent limitations LLMs face when addressing specialized questions and context-dependent domains. Applied to national security, these models can help support campaign planning, wargaming, and studying escalation dynamics (Table 1).

▲ Table 1: Training Datasets. Source: CSIS Futures Lab.

First, the authors had the model summarize common criticisms of the UCP to provide a baseline leveraging 119 prior studies of UCP reform. Of note, much of the criticism matched the prevailing debate, which is not a surprise since LLMs excel at basic synthesis.

Second, the authors used this baseline to explore alternative UCP reform options weighted against different combinations of datasets to, in essence, force productive hallucinations and novel insights. These datasets included 70 documents on Chinese military doctrine and competitive strategy; 205 documents on deterrence theory, military history, and operational art; and 232 document excerpts on international relations cutting across multiple theoretical schools of thought.

▲ Figure 1: Model Threads.

Using these datasets, the authors asked the model (1) to create an illustrative scenario depicting global competition between the United States and China and (2) to recommend a new set of combatant and functional commands based on the scenario that ensures the United States is sufficiently aligned to counter the Chinese Communist Party (CCP).

The authors ran this sequence four times to explore where the model returns converged and where they diverged (Figure 1). In model one — “The Hunt for Asymmetries” — the authors used concepts from competitive strategy and assessments of the People’s Liberation Army (PLA) military strategy to analyze UCP rewrite options. In model two — “The Rational Past as Prologue” — the authors used deterrence theory, military history, and concepts from bargaining theory to analyze UCP rewrite options. In model three — “Anarchy Shapes All Decisionmaking” — the authors used texts on realism from international relations theory. Last, in model four — “Stories Guide the Art of the Possible” — the authors used a mix of social constructivism and critical perspectives to analyze UCP rewrite options. Below the authors summarize the results from each iteration.

Last, to help visualize and describe the alternative strategic logics, the research team tailored prompts to produce stylized paintings using the Midjourney AI Pro art generator. To create a common reference point, the team combined international relations theory references with prompts requesting the LLM to produce a “painting in the style of Norman Rockwell.” The use of this famous American artist situates the visual logic in classic portrayals of American nostalgia, a deliberate effort to show alternatives to the Cold War history of the UCP for a new era of great power competition.

The Hunt for Asymmetries

The first thread focused on using net assessment and documents on Chinese military strategy to assess the UCP. The thread depicted an outsize role for U.S. Indo-Pacific Command (USINDOPACOM) supported by existing regional commands — consistent with current approaches to global integration — and a host of new functional commands. The thread discussed the need to organize roles and missions around an evolving assessment of regional military balances relative to force posture. It also discussed the need to evaluate military balance and force posture in a search for asymmetry. In other words, net assessment should drive competitive strategy and alteration of assigned forces within the existing UCP structure and the current global integration framework, placing USINDOPACOM as the global coordinating authority for competition with China, supported by regional and functional commands.

The thread sought to expand the number and type of functional commands required to support long-term competition with China. First, it expanded U.S. Cyber Command to include a larger role in technology policy and research, development, testing, and engineering. In addition to coordinating cyber offense and defense activities, the thread detailed new responsibilities for protecting U.S. technology from Chinese cyber theft and focusing investments in advanced military technology, currently overseen by the undersecretary of defense for research and engineering as well as service-level entities. Second, the model discussed the need for a new functional command focused on coordinating intelligence, surveillance, and reconnaissance activities globally, with an eye toward countering Chinese battle networks and developing indicators and warnings on military capabilities and troop deployments.

Interestingly, the tailored LLM suggested the United States should, through the UCP, get its act together when it comes to nonmilitary aspects of national security policy. First, the model discussed the need for a new functional command focused on economic security. The interagency command would counter Chinese economic strategies harmful to U.S. interests, counter economic coercion against partners and allies, and coordinate interagency measures addressing issues such as trade imbalances, intellectual property protection, and other economic security concerns like the labor market. Second, the model proposed new structures dedicated to human assistance and disaster relief to promote stability and enhance the reputation of the United States in regions where China is expanding its influence. These new soft power approaches are intended to appeal to the hearts and minds that China has courted in the past or is actively courting. Examples include countries in the Belt and Road Initiative or regions like Africa and Latin America.

Much of the discussion is consistent with the large body of debates on interagency reform and past tendencies to expand, vice optimize, the UCP to solve evolving strategy challenges. What is interesting, however, is the model’s conclusion that integrating civilian-military effects into an operational and operational-level command structure is necessary to gain a long-term advantage against the CCP.

▲ Figure 2: Reimagining the UCP Using Net Assessment. Source: Benjamin Jensen. Midjourney Pro Prompt: “Andrew Marshall net assessment explaining Cold War strategy painting in the style of Norman Rockwell.”

The Rational Past as Prologue

The second model combined documents on deterrence, military history, and operational art. Of note, the scenario the model returned anchored global competition around a potential crisis in the South China Sea, drawing on principles from deterrence theory more than from operational art or military history. Most of the output discussed using diplomacy alongside demonstrations of military capability with only a limited discussion of combined arms and lines of operation. In other words, escalation management loomed larger than gaining a near-term decisive military advantage.

At the same time, the model envisioned competition with China as global and cutting across multiple domains in an effort to support a cost-imposition strategy. The scenario discussed using security cooperation in Africa to counter Chinese access and diplomacy to counter growing Chinese influence in Eastern Europe. The scenario discussed working with partners and the private sector to make it more difficult for China to gain an advantage in space and cyberspace. Last, the scenario focused on diplomacy and economic sanctions as the main effort, seeking to signal to China the risks of continued escalation. According to the model, these combined lines of effort created a more synchronized approach to strategy:

“All these activities have a constant backdrop of political diplomacy, economic negotiations, and information campaigns to create a comprehensive response strategy. These strategies collectively work towards deterring China from aggressive action and work to maintain the balance of power in line with international rules and norms. This kind of strategizing effectively uses the UCP and its multiple combatant commands, integrating operational art and deterrence theory on a global scale.”

Based on this strategy, the model recommended the following changes to the UCP. First, it advocated consolidating existing combatant command structures. U.S. European Command (USEUCOM) and U.S. Central Command (CENTCOM) would be merged to form a new Eurasian Command to better link global competition with Russia and China outside of USINDOPACOM. The same logic applied to U.S. Northern Command (USNORTHCOM) and U.S. Southern Command (USSOUTHCOM). The model recommended a new Americas command merging the two to better prioritize countering China’s growing influence in South America. The model also recommended merging several of the functional commands, including the U.S. Cyber Command and U.S. Space Command, to better synchronize multidomain operations.

Second, the model recommended an operational command structure similar to the net assessment thread (“The Hunt for Asymmetries”) to make economic competition a central pillar of national strategy. Of note, the model saw the primary role of the functional command as coordinating interagency activities to counter the Belt and Road Initiative and Beijing’s economic priorities. The model identified the need to coordinate nonmilitary forms of power with traditional military roles and missions to manage long-term competition with China.

▲ Figure 3: Reimagining the UCP Using Rational Deterrence. Source: Benjamin Jensen. Midjourney Pro Prompt: “Thomas Schelling explaining nuclear deterrence using game theory on a chalk board painting in the style of Norman Rockwell.”

Anarchy Shapes All Decisionmaking

When weighted with realist texts, the model focused more on how best to manage escalation regardless of the structure of combatant commands alongside the need to better integrate different agencies and instruments of power. Specifically, the model highlighted war control — a PLA concept for controlling the pace and intensity of escalation — as providing the United States an opportunity to take a measured approach to maximizing its security. Second, the model saw the future of integrated deterrence less in technology and more in how well the United States integrates multiple instruments of power and works alongside partners and allies. Last, the model saw the future as a global struggle and highlighted the need for the United States to focus on maintaining favorable regional balances of power.

▲ Figure 4: Reimagining the UCP Using Realism. Source: Benjamin Jensen. Midjourney Pro Prompt: “Machiavelli discussing political realism painting in the style of Norman Rockwell.”

Stories Guide the Art of the Possible

When weighted with constructivist and critical theory texts, the model prioritized norms, power dynamics, and underlying narratives. This focus meant the analysis of UCP options prioritized shaping stories and how people perceive the word over regional balances of power. The model focused on a new global engagement command with elements similar to those discussed in the operational art and deterrence theory iterations, albeit based on different assumptions. Modern competition was seen as a war over ideas. In this logic, the United States needs to address not just China’s current actions but also the wider power dynamics and structural issues at play in foreign policy and how they collide in stories, propaganda, and global influence campaigns. The model assessed that these responses could involve challenging dominant narratives around sovereignty, power, and security but should avoid imposing Western-centric narratives and norms in diplomatic initiatives. Like previous threads, the insights support new, novel structures linked to coordinating nonmilitary instruments of power.

▲ Figure 5: Reimagining the UCP Using Constructivism. Source: Benjamin Jensen. Midjourney Pro Prompt: “a feminist criticizing grand strategy painting in the style of Norman Rockwell.”

Policy Options to Explore

While all statistical models are incomplete, some are useful, and generative artificial intelligence (AI) is no different. The ideas that emerged from this experiment suggest three options worth further deliberation for rewriting the UCP.

Enhance Command Structure for Global Engagement and Competition

Across the threads, the generative model highlighted a need to better organize the DOD — if not the entire national security enterprise — for competition short of war. These campaigns would combine traditional security cooperation activity with new efforts to protect U.S. technology and economic access. Better still, this realignment could help better connect competition and contingency planning, a perennial challenge in the Pentagon.

The challenge in the current UCP design is that the authorities for these activities are divided between multiple government agencies that do not have similar regional planning frameworks (e.g., DOS and DOD). Furthermore, these activities are not strictly military, making them outside the purvey of geographic combatant commands. Getting better synchronization, therefore, would likely require deeper debate about how to design a national security system that currently contrasts DOD efforts to manage regional security challenges with the country-specific orientation of the DOS. That is a tall order and one likely to involve Congress.

In the near term, military planners should consider using the UCP to task each combatant command to stand up interagency joint task forces for competition. This standing task force would enact current theater and global competition plans but serve as a focal point for better coordinating with other agency activities. In the absence of new authorities, a large amount of the task force’s time would be used to establish interagency awareness and develop a common understanding of the range of ongoing programs and activities in the region and how best to align competition planning with them. This interagency awareness over time could evolve to inform national security coordination through the NSC and augment experiments with battle networks currently ongoing through the Global Information Dominance Experiments (GIDE) managed by the Chief Digital and Artificial Intelligence Office (CDAO).

Merge Space, Cyber, and Information Effects

Across the threads, an approach to military strategy emerged that stands in stark contrast to the last decade of UCP rewrites. Rather than continue to create new functional commands, the model proposed fusing them into a single command somewhat similar to the recently disbanded PLA Strategic Support Force. In some respects, such an entity could even be folded back into U.S. Strategic Command for a one-stop shop for deterrence and setting conditions for modern war fighting. The operative question is what the DOD gains from creating multiple functional commands instead of streamlining and rationalizing them into new entities.

The UCP rewrite team could consider incremental options, such as focusing on global competition. For example, it could experiment with establishing a sub–unified command in each geographic command focused on integrating space, cyber, and information effects into ongoing operations. While interesting, this incremental approach would likely create duplicate structures and would not address the pull to create four-star headquarters for each new domain alongside new services. Clearly, modern competition and war fighting require the ability to integrate multiple domains. Therefore, assigning roles and missions with this unity of effort in mind should drive UCP reform.

New Regions, Fewer Components

Last, the model suggested realigning existing geographic combatant command boundaries. The model played with new seams for global competition with China, including an Americas Command merging USSOUTHCOM and USNORTHCOM and a Eurasia Command merging USEUCOM and CENTCOM. This logic is similar to that of earlier experiments during the Cold War that went as far as creating one command for overseas forces and another for those in the United States.

Left unsaid by the model was that merging existing geographic combatant command boundaries could produce cost savings and provide a model for even more radical reductions. For example, are service component commands truly necessary? In both his planning guidance and subsequent comments, General Berger, the 38th commandant of the Marine Corps, expressed concerns about a lack of naval integration that inadvertently emerged as a result of the Goldwater-Nichols Act. Maybe it is time to reduce the number of service components at each combatant command, if not eliminate them outright, since service headquarters can coordinate directly to assign forces.

The current team working on the UCP could explore opportunities to reduce the number of headquarters and build structures that better align with the shape of global competition with the CCP.

Conclusion: Back to the Ritual

UCP rewrites are both an opportunity and a bureaucratic conundrum for those who oversee the ritual. Wasting that opportunity, despite the painful debates that need to occur, would abdicate professional responsibility both for those in the fight and for those watching from the outside but still dedicating themselves to national security reform.

The contributions of this report are illustrative, not conclusive. The series of LLMs tailored with key documents highlight big ideas but nothing especially new. For years, national security practitioners have known the current design of the U.S. national security enterprise is broken. The UCP is just one part of that larger Cold War–era bureaucracy in need of a reboot.

The models reiterate the need to rethink functional and geographic command structures alongside their roles and missions. If nothing else, there are significant cost savings to be had from merging commands as well as rethinking the current map. Staffers in the bureaucratic fight over the next UCP must at least answer the question, Why shouldn’t the United States merge the U.S. Space and Cyber Commands? More interesting is the need to imagine new ways of integrating military and nonmilitary power to manage long-term competition with the CCP. Again, ongoing dialogue inside the Pentagon should at least consider an Americas and Eurasia Command and, more importantly, what command structures would best support managing nonmilitary competition with China globally and bridging interagency divides.

Outside of the models, many of these ideas need not take the form of entirely new four-star commands. Rather, new interagency joint task forces — including novel approaches like Joint Task Force Ares, used to combat the Islamic State and subordinate unified commands — offer flexible structures for managing a complex, multidomain, multi-instrument global competition. Even more radical would be using a congressional commission to explore rewriting key authorities and legal code linked to how the United States organizes for national defense and long-term competition. Regardless of what form the next UCP takes, the process of building it should ask hard questions and demand more than answers forged by bureaucratic path dependence, satisficing, and proponency. The world is too dangerous to avoid hard choices; just ask our large-language model.

Benjamin Jensen is a professor in the School of Advanced Warfighting at the Marine Corps University and a senior fellow in the Futures Lab at the Center for Strategic and International Studies (CSIS) in Washington, D.C.

Kathleen McInnis is a senior fellow in the International Security Program and the director of the Smart Women, Smart Power Initiative at CSIS.

Jose M. Macias III is a research associate in the Futures Lab at CSIS and a Pearson fellow at the Pearson Institute for the Study and Resolution of Global Conflicts at the University of Chicago.