Autonomous Systems, Recoverable Data: AI Has Changed the Security Equation for the Battlefield Edge
Last updated 08:55 AM EDT • Read time: 6m
The Department of War is undergoing one of the most significant transformations in military operations. Artificial intelligence is moving decision-making to the tactical edge. Autonomous systems are proliferating across every domain. Data is becoming as decisive as firepower. And military planners increasingly envision a future where thousands of unmanned and autonomous systems operate simultaneously across highly contested battlespaces. This transformation is reshaping not only how wars will be fought, but also where America's most valuable military information resides. Increasingly, it resides at the edge.
Artificial intelligence has fundamentally changed the architecture of military computing. Every AI-enabled unmanned aircraft, autonomous ground vehicle, maritime platform, ISR sensor, edge compute node, and autonomous command-and-control system now processes, stores, and acts on operational data locally rather than relying on centralized command centers. As AI pushes computing to the tactical edge, battlefield platforms have become repositories of some of the military's most valuable information.
But what happens when the platform does not return? Whether it is an inexpensive quadcopter, an MQ-9 Reaper, an autonomous surface vessel, or a collaborative combat aircraft, physical loss is an operational reality. The platform may be expendable, but the data certainly is not. Recent conflicts including the loss of MQ-9 Reapers during operations involving Iran and its regional proxies demonstrate that even the most sophisticated unmanned systems remain vulnerable to capture or exploitation.
What an adversary gains extends far beyond sensor imagery. Captured systems may reveal mission plans, target packages, navigation routes, flight histories, communications configurations, electronic warfare parameters, cryptographic credentials, and network architecture information that can compromise future operations and expose tactics, techniques, and procedures. More valuable still are the AI models, proprietary algorithms, software, and firmware that enable autonomous navigation, sensor fusion, target recognition, and electronic warfare resilience. These capabilities represent years of research and billions of dollars of investment, yet if recovered intact they provide adversaries an opportunity to analyze, replicate, develop countermeasures against, or even accelerate competing capabilities.
The risk extends well beyond the vehicle itself. Every unmanned platform depends on an ecosystem of mission-planning workstations, ground control stations, ruggedized laptops, tactical servers, portable SSDs, maintenance systems, and edge computing devices that collectively contain an even more complete picture of military operations. Compromising one platform may expose a mission; compromising its supporting ecosystem can reveal how an entire force plans, communicates, and fights.
Ukraine has demonstrated this repeatedly. Downed systems are routinely recovered and exploited, with the intelligence extracted often proving more valuable than the hardware itself. As AI continues moving intelligence and decision-making to the tactical edge, the strategic value of the data will increasingly exceed that of the platform carrying it. In the emerging era of AI warfare, losing a trained model, mission dataset, or autonomous software stack may prove far more consequential than losing the vehicle itself.
This changes how the Department of War should think about survivability. For decades, survivability has largely focused on protecting the platform - stealth, mobility, armor, countermeasures, and redundancy. Tomorrow's force must place equal emphasis on protecting the information.
Every autonomous platform should be designed with the assumption that physical capture is not merely possible but expected. Sensitive information should remain protected even when the platform does not. Mission data should remain inaccessible. Cryptographic material should resist extraction. AI models should be protected against compromise. Intellectual property should remain secure. Operational data should never become an intelligence gift simply because a vehicle failed to return.
These protections cannot be limited to the aircraft, vehicle, or vessel itself. They must extend across the entire operational ecosystem that supports autonomous missions from mission planning systems and portable storage devices to edge servers, maintenance equipment, and tactical computing platforms that collectively enable autonomous operations.
The Department of War has correctly recognized that future conflict requires mass, autonomy, and artificial intelligence. It has invested heavily in accelerating production, expanding the industrial base, and fielding new generations of autonomous capability.
Now doctrine must evolve once again. The discussion should no longer be limited to how many autonomous systems can be built or how inexpensive they can become. It should ask a more enduring question: When the platform is inevitably lost, what happens to everything it knows?
The answer will increasingly determine not only the outcome of a single mission, but whether years of operational advantage, technological innovation, and strategic investment remain America's or become an adversary's.
As artificial intelligence pushes computing, decision-making, and operational intelligence to the tactical edge, protecting battlefield data can no longer be viewed as an IT security problem. It has become a fundamental requirement for operational survivability.
The good news is that data compromise following platform loss is not inevitable. Data-at-rest protection is a mature discipline built on well-established cryptographic and authentication technologies designed specifically to prevent unauthorized access when a device is lost, stolen, or captured. Because battlefield systems frequently operate disconnected from enterprise networks, these protections must reside within the endpoint itself rather than depend upon network-based security controls.
For systems handling classified information, the National Security Agency's Commercial Solutions for Classified (CSfC) Data-at-Rest architecture provides a proven framework built around independent layers of authentication and encryption. When implemented using NSA-approved architectures and components, these layered protections significantly reduce the risk that an adversary can access sensitive information even after gaining physical possession of the device.
The risk is real. It is growing as AI and autonomous systems proliferate across the battlefield. Fortunately, the engineering solutions already exist. The remaining challenge is ensuring that protection of data at rest becomes as fundamental to future force design as mobility, survivability, autonomy, and communications. The United States has accepted the doctrine that some platforms will be lost. It should never accept the doctrine that the data they carry must be lost with them.
Brett Hansen, Cigent CEO, has 30 years of IT experience. Prior to joining Cigent, held leadership positions with companies providing SaaS data solutions. He was also an executive at Dell for over a decade overseeing Client Software organization including developing and leading Dell Data Security business that provided integrated hardware and software security solutions. Brett started his career with IBM holding leadership positions with various software organizations focusing on IT management, development, and security.