When the world's at stake,
go beyond the headlines.

Popular culture often distorts autonomy into science fiction caricatures. This framing obscures the real challenges and opportunities facing the U.S. military: autonomy as a layered, incremental capability, always shaped by mission context.

Ukraine’s “Operation Spiderweb,” in which over a hundred drones conducted a coordinated, semi-autonomous attack, illustrates this reality. Operators tasked and launched the systems, but autonomy enabled distributed navigation, deconfliction, and timing. It wasn’t a leap to science fiction. It was a practical demonstration of how autonomy can multiply effects and impose asymmetric costs.

For the Pentagon to realize a commercial-first and more autonomous force, it must overcome a lack of clarity around how the defense industrial base defines autonomy. Some use the term when referring to facets of an augmented operator experience, others characterize it as the “end game” like Skynet from The Terminator, and sales-driven vendors eagerly add “autonomous” to anything from advanced autopilot features to fully independent systems.

As leaders in the private sector applying advanced technology to the Defense Department’s autonomy missions, we have a commercial interest in delivering autonomous capabilities to warfighters. But at this crucial stage in autonomy adoption, we believe the United States can accelerate progress by establishing shared definitions, standardizing key technical layers to enable interoperability, and building operator trust through education and early integration.

Autonomy in Action

Within the defense sector, there are numerous organizations and partners, each with its own portfolio of platforms, that the Pentagon would ideally want to communicate with. Interoperability, or the ability for systems to exchange and make use of information, is incredibly important for autonomous systems to work together successfully and achieve common goals. But the systems will not be able to effectively communicate without a shared taxonomy and common language. Without shared definitions, the government cannot make informed investments and choices when it comes to the partners it selects.

Take, for example, protecting the Taiwan Strait, a potential scenario looming large. In such a scenario, the Air Force, Navy, and Marine Corps would all need to be in position to communicate and defend Taiwan in the event of a repeat crisis. The problem? Individual services are pursuing their own approaches to autonomy for airborne, maritime, and land-based systems. If those systems cannot communicate at the point of crisis, it would be the equivalent of radios that cannot transmit across combat fields. The consequences could be perilous.

The Army’s Low Altitude Stalking and Strike Ordnance program is another critical example. The program is poised to advance modern warfare by delivering portable, precise, and lethal unmanned munitions directly to infantry units — yet it also underscores nuances of autonomy. The program operationalizes autonomous systems that can loiter, track moving targets using onboard sensors, and strike with lethal precision. A human still selects the target and confirms the engagement in most operational concepts. By communicating outcomes rather than means, industry can align capability and investment towards the Army’s objectives. By using a clear and standardized vocabulary, military planners and policymakers understand these systems, their benefits, and their limitations as they are fielded.

Navigating Standards

It can be challenging to navigate the ambiguity between autonomy layers when integrating across partners, something that is currently taking up considerable human effort.

Government reference architectures are designed to solve this problem. These are standardized frameworks created by government agencies to guide the design, implementation, and integration of systems. For example, the Agile Mission Systems Government Reference Architecture was an attempt to enable quick interoperability by integrating payloads. It has been extended to be used for theater-level operations between different capabilities, and most recently, autonomy. This is a step in the right direction, but there is a clear need to move faster. And worse, each service still adheres to its own frameworks, making coordination between them difficult. Add international partners with their own global standards into the mix, and the complexity heightens.

The key is targeting the right layer, or layers, for standardization and applying it in a common way to all relevant systems. The internet has been successful for all these years by standardizing transport protocols. The Air Force’s Universal Command and Control Interface, by contrast, targets messaging protocols, which is an example of a different approach to layering. A strategy for successful autonomy will be similar: Find one layer where everyone agrees on definitions and communication, leaving room for industry innovation above and below.

The layer of standardization also shapes acquisition, providing the interfaces for competing products or integration with many different suppliers. Industry innovation to capture government funding will evolve around the standards used in acquisition. The more acquisition is built on the standardized layer, the more industry will adopt, optimize, and evolve its products. Where each service defines its own standards and layers, the acquisition pool is diluted, and industry solutions evolve more slowly and benefit smaller communities of interest. An intentional department-wide approach to autonomy acquisition that uses standards to shape the market will accelerate the evolution of advanced autonomy operational capabilities.

Consider, for example, the rise of the Universal Serial Bus, or USB. Before the USB’s release in 1996, users had to pry open their computers and install hardware themselves to get the communication port they needed to connect devices like a printer or scanner. A lack of standard practices among the many suppliers in the industry made this a challenging task, even for skilled professionals. Yet within a few years, the USB had become the default connection standard for many peripherals, ultimately becoming an industry standard while making legacy technology like the floppy disk obsolete.

What can defense technologists learn from the USB’s success? First, USB-enabled devices from various manufacturers needed to work together by creating a common standard for data transfer and power delivery. More importantly, the USB highlights how standardization can unify fragmented systems and improve interoperability while also making technology more accessible for end users and contributing to accelerated adoption.

Building Operator Trust

Beyond taxonomy and standards, a successful autonomy framework should be grounded in operational needs and built on operator trust in the system. The stakes of deploying autonomous capabilities in contested environments are very high, so, understandably, operators approach these systems from a conservative and cautious point of view.

Calibrated trust is essential. From the very beginning, operators should be educated on the systems and what they do. Early integration of the technology into familiar workflows is critical in allowing them to see its practical use and begin building trust. Taking this a step further, it is important to give operators tools to analyze the behaviors of these technologies and run diagnostics to further understand how these systems behave.

Each domain brings unique experiences and influences to the table, and it can be a challenge to ensure each is heard and understood. It is important that the vendors of these autonomous systems are engaging with each of these communities, embracing the knowledge share, and ensuring these systems will work for their individual problem sets.

The Policy Imperative

There are complex nuances in achieving autonomy, from standards to taxonomy to trust. But the strategic imperative is clear: If the Pentagon does not lead in setting a common language for autonomy, industry marketing will fill the void, sowing confusion across the force.

The Pentagon should move quickly to establish a standard layer for autonomy messaging and tasking, much as transport protocols unified the internet, or risk incoherence in future conflict. The fog of marketing could thicken into a fog of war, and that is a risk the United States cannot afford.

Randy Yamada is a vice president within Booz Allen’s Defense Technology Group, overseeing autonomy and embodiments of physical AI.

Tom Schaefer is a vice president of Hivemind Engineering at Shield Al, where he leads the development of the company’s suite of autonomy platform products for software developers and organizations building, testing, or evaluating autonomy for intelligent machines.

Image: U.S. Army Intelligence Center of Excellence via DVIDS.