Bryan Clark testifies before the House Armed Services Committee.
Chairman Gallagher, Ranking Member Khanna, and distinguished members of the committee, thank you for the opportunity to discuss the Replicator initiative and the Department of Defense’s (DoD) overall efforts to field uncrewed systems. Since its announcement last month, Pentagon officials have said little about Replicator, except that it is intended to field thousands of uncrewed systems during the next two years. The initiative’s main goal, according to Deputy Defense Secretary Hicks, is to provide attritable mass that can counter the geographic and capacity advantages enjoyed by China’s People’s Liberation Army (PLA) in potential Western Pacific confrontations.
There is little evidence the DoD and its industry partners can field thousands of operationally relevant uncrewed systems in the next two years. The Pentagon’s anemic procurement of uncrewed systems has generally discouraged industry from ramping up its production capacity. And in those cases where privately funded companies or traditional defense contractors have invested in manufacturing infrastructure, they have lost money or exited the sector entirely.
But production capacity is not the biggest problem. Even if Replicator is successful, simply adding mass to today’s US military is unlikely to improve its ability to deter or defeat China. With its proximity to likely areas of conflict, lack of global responsibilities, and ability to focus on US forces, the PLA can field targets at a lower cost and greater scale than the US military can generate successful shots on target. If it competes only in terms of mass, the DoD will find itself perpetually playing catch-up.
However, there are glimmers of hope. In her discussion of Replicator, Deputy Secretary of Defense Hicks suggested the initiative is designed to exploit the creativity of US warfighters in addressing problems faced by today’s operational commanders. Compared to chasing mass, this approach offers a better path to gaining advantage over the PLA and could mitigate the challenges US industry will likely face in rapidly producing thousands of militarily relevant uncrewed systems. But unlocking Replicator’s ability to deliver innovative solutions for pressing operational problems will require the DoD to integrate uncrewed systems into the mainline force rather than continuing to treat them only as surveillance systems or extensions of crewed ships and aircraft.
Until the twenty-first century, the DoD mainly accomplished integration by aligning doctrine and procedures because humans operated nearly all equipment. Today, automation and machine-to-machine communication between vehicles, platforms, and systems reduces the need for human operators to act as intermediaries—in most missions, increasing operator involvement is more likely to reduce performance than to improve it.
Integration, rather than long-term research and development (R&D), should therefore be the Pentagon’s focus for Replicator. The DoD does not have time to develop sophisticated new uncrewed systems from scratch, nor does it need to. Existing and emerging uncrewed technologies can give US and allied militaries the edge they need against the PLA if combined with existing units and orchestrated in ways that create adaptability for friendly forces and uncertainty for the enemy.
Replicator Should Pursue Adaptability, Not Mass
In prioritizing mass, Replicator seems to rely on the same uncrewed system characteristics US forces have exploited for decades, summarized in figure 1. Because they do not carry human operators, even expensive uncrewed vehicles may be lost to combat or equipment failure with little regret. Without the confines of human limitations, uncrewed systems can operate in unforgiving environments or circumstances such as space. And without human operators, uncrewed vehicles can be less expensive than their manned counterparts due to fewer requirements for life support, protection, live training, or multi-mission capability.
Figure 1: System-Level Value Proposition for Pursuing Uncrewed Solutions
High-priority uncrewed systems being developed by the DoD like the Air Force’s Autonomous Combat Platform (ACP, formerly Collaborative Combat Aircraft), Army Air-Launched Effects (ALE), and Navy Large Uncrewed Surface Vessel (LUSV) are exploiting these characteristics to extend the reach and persistence of their crewed ship, aircraft, or artillery teammates. Such manned-unmanned teaming (MUM-T) made sense when the US military was dominant and trying to maximize efficiencies. However, this approach also tends to perpetuate the limitations of crewed systems, which operate in standardized formations for sustainment and protection and rely on well-defined doctrine and procedures that facilitate training.
Binding uncrewed systems to the predictable operations of their crewed counterparts plays into the PLA’s concept of system destruction warfare, or systems warfare. Under this approach, the PLA assesses the systems of systems (SoS) US forces are likely to use in combat and their potential vulnerabilities. The PLA then develops and fields capabilities that can attack what it perceives as US weaknesses and undermine the ability of US and allied militaries to intervene on behalf of allies like Taiwan. For example, the PLA fields a variety of electronic warfare systems that target key US networks and has exploited commercial and proliferated military technologies to undermine traditional US advantages in air defense, precision strike, and long-range power projection.
The US military will need to be less predictable and more adaptable to gain an edge against the PLA. Replicator could help if it prioritizes operational innovation, as Secretary Hicks noted in her announcements, and seeks advantages from force employment and associated command, control, and communications (C3) capabilities instead of strictly through superior weapon, sensor, or platform technology. This, rather than mass, is the goal of the Ukrainian military in fielding uncrewed systems, whose efforts Secretary Hicks cited in her Replicator announcement. Uncrewed systems on the sea and in the air have provided Ukraine’s military modes of attack and fires coordination that Russian forces have often been unable to anticipate or counter.
The US military is already pursuing more adaptable operational approaches that could employ systems emerging from the Replicator initiative. The Joint Warfighting Concept (JWC), Distributed Maritime Operations (DMO) concept, and Joint A ll-Domain Command and Control (JADC2) initiative rely on distribution; recomposable force packages; and long-range effects chains connecting sensors, commanders, and weapons or electronic warfare systems to undermine PLA systems warfare. By degrading an opponent’s sensing and sense-making while affording US forces more options for offensive action, these initiatives aim to increase the US military’s lethality and resilience.
Budget constraints will prevent the US military from becoming more distributed and recomposable by simply growing the existing, mostly crewed, force. To surmount this obstacle, the DoD will need to expand the proportion of the force that is uncrewed while investing in the ability to identify and integrate new effects chains using AI-enabled C3 software. Rather than acting as extensions of crewed units, uncrewed systems in future effects chains will need to perform as independent elements of force packages or SoS.
As figure 2 summarizes, adopting an SoS approach to force employment will allow the US military to fully exploit the characteristics of uncrewed systems. Because they are less expensive compared to crewed units, uncrewed systems can enable scaling the force to increase distribution. The advent of a robust commercial robotics technology ecosystem further expands this opportunity by lowering costs and avoiding time-consuming R&D. With their scale and expendability, uncrewed systems can expand the variety of effects chains available to commanders and the dilemmas they impose on adversaries—provided forces treat them as independent players in an SoS. And because uncrewed systems can be more specialized and modular compared to crewed units that require multi-mission capability, forces can more easily plug them into effects chains to adapt an SoS to new missions or environments.
Figure 2: Systems-of-Systems Value Proposition for Replicator
A useful comparison is Australia, which faces similar military challenges from the PLA as the United States and is pursuing a range of uncrewed technologies, some of which are encompassed under the Australia-United Kingdom-United States (AUKUS) agreement’s Pillar Two. Hudson Institute is working with the Australia Department of Defence (ADoD) to refine its efforts at fielding uncrewed systems. Because it lacks the resources of the US DoD, the ADoD has had to scope its uncrewed system development to emphasize relatively mature technologies and near-term challenges faced in its near-abroad, such as defending Australia’s northern approaches from intrusion or attack. Perhaps the most innovative dimension of Australia’s uncrewed system development is its equal promotion of opportunities to gain asymmetric advantage, rather than simply fill gaps in current capabilities. By using uncrewed systems to open up new concepts, the ADoD seems to exemplify the characteristics needed in the Replicator initiative.
Replicator Should Align Requirements with Limits on Autonomy
Adopting a SoS context in Replicator will also help field uncrewed systems more quickly, because together the elements of a SoS can mitigate limitations in uncrewed system autonomy, or the degree to which a system can be self-governing or operate without outside support in executing a task or function. A popular characterization of uncrewed systems is that they are “autonomous,” but this is an overstatement because uncrewed systems depend on other force elements for essential support functions, from navigation to logistics.
Crewed ships, aircraft, or ground vehicles are also limited in their autonomy, but the constraints on uncrewed system autonomy are often more severe in degree depending on the mission, operating environment, and sophistication of the unit in question, as shown in figure 3. Some combinations of complexity and duration will be unachievable with available uncrewed vehicle technology, as indicated in the figure’s upper right. Systems that need to operate in this region, including several of the DoD’s high-profile uncrewed vehicle programs, drive R&D efforts that take years to culminate.
Although hardware will define the upper limit of how many variables a machine like an uncrewed vehicle seeks to control (e.g., how many control loops or similar control logic implementations are in its hardware and software), the use case dictates the number of necessary control loops. For example, driverless automobiles that operate outside known environments require a very large number of controlled variables, and with current technology they cannot operate for long periods without operator intervention, as recent accidents suggest. So for now driverless automobiles can only conduct short trips in environments like urban centers that are well mapped and where the vehicles can gather large amounts of data regarding local traffic patterns and behaviors.
Figure 3: Relationship between Vehicle Sophistication and Endurance (SUUV: Small uncrewed undersea vehicle; ASW: Anti-submarine warfare)