The AI Co-Pilot for theEMS Battlefield

A new generation of real-time AI tools now gives operational commanders what they have never had before:

AI-ENABLED ELECTROMAGNETIC SPECTRUM OPERATIONS • COMMANDER’S DECISION SUPPORT

A new generation of real-time AI tools now gives operational commanders what they have never had before: continuous machine-speed monitoring of the electromagnetic environment, intelligent threat attribution, and tactically reasoned recommendations for countering jamming and operating through degraded communications.

 • February 2026

BLUF — Bottom Line Up Front

The Army is fielding and prototyping a layered AI-enabled system that, for the first time, gives the operational commander a real-time picture of the electromagnetic battlespace, automated threat identification, and machine-generated courses of action for degraded operations. The key components — S2AS, TLS BCT Manpack, EWPMT-X, and the NGC2 common data layer with ML triage — are either in rapid fielding or scheduled for FY2026 demonstration and procurement. At the joint level, Palantir’s EMBM-J DS prototype brings AI-assisted COA analysis directly to joint electromagnetic spectrum operations planners. The overarching framework is the Army’s Next Generation C2 initiative, whose explicit design goal is enabling human decisions at machine speed in a denied, degraded, intermittent, and limited (DDIL) environment. The capability exists — the challenge is integration, classification-level interoperability, and building the command culture to act on AI recommendations without waiting for certainty that will never come.

THE COMMANDER’S PROBLEM

Until very recently, the operational commander’s awareness of the electromagnetic environment was a manual, intermittent, specialist-dependent process. The CEMA officer brought a picture to the commander during the intelligence update brief. The spectrum manager worked from spreadsheets. Jamming was identified when communications failed — after the fact, by inference, not by detection. The OPFOR at NTC Fort Irwin described the result precisely: Blue Force commanders, deprived of their common operating picture, froze. They did not know what was happening in the spectrum, could not distinguish jamming from equipment failure or interference, and could not make a confident decision about whether to fight through degradation or transition to an alternate communications tier.

The problem has three distinct layers. The first is monitoring: the commander needs a continuous, visual, machine-speed picture of the electromagnetic operating environment (EMOE) — what signals are present, whose they are, what the friendly emission signature looks like, and where anomalies suggesting jamming, spoofing, or adversary reconnaissance are appearing. The second is threat detection and attribution: the system must identify not just that something is happening in the spectrum, but what it is, who is likely doing it, and what operational intent it signals. The third is decision support: given a characterized threat and a current mission, the system must recommend actionable responses — maneuver options, communication tier changes, electronic attack targeting, or emission control measures — and express those recommendations in terms the commander can act on without a specialist translator.

Each of these layers now has a corresponding Army program in development or fielding. None of them is fully mature. But for the first time, the architecture to solve all three exists, is under test, and is on a fielding timeline. Understanding what each layer does — and where the current gaps are — is essential for any analysis of where AI truly transforms the commander’s EMS decision calculus.

LAYER 1 — MONITORING: S2AS AND THE REAL-TIME SPECTRUM PICTURE

The Spectrum Situational Awareness System (S2AS) is the foundational monitoring tool. Developed by Product Manager Electronic Warfare Integration under PEO IEW&S and built on mature commercial off-the-shelf technology, S2AS provides commanders with a real-time electromagnetic spectrum situational awareness display integrated directly into the command post’s C2 environment. It senses, detects, and reports emitter activity across the battlespace, visualizes friendly signal signatures — including those a unit may not know it is emitting — and identifies sources of electromagnetic interference, whether from coalition equipment, inadvertent fratricide jamming, or adversary systems.

The system is designed to work in conjunction with EWPMT-X, the Army’s next-generation electromagnetic warfare planning and management tool, so that what S2AS senses flows directly into the planning tool the CEMA cell and commander use to make decisions. Maj. Cedric Harris, assistant product manager for EW Integration, described the combined effect: “Modernizing spectrum awareness capabilities will enhance commanders’ ability to visualize and control the electromagnetic environment, enabling more effective decision making and improving protection to win the fight for spectrum dominance.”

S2AS completed a three-week User Assessment in August 2025 with soldiers from the 25th Infantry Division, 101st Airborne Division, 11th Airborne Division, and Special Operations Command — units across the operational spectrum, not just EW specialists. An operational demonstration test event was planned for Q1 FY2026, with a rapid fielding decision targeted for Q2 FY2026. The system is available in handheld and vehicle-adaptable configurations, meaning spectrum awareness can extend from the brigade command post down to dismounted company-level operations. This is the first time organic spectrum visualization has been available to maneuver formations at that echelon.

Critically, S2AS addresses one of the most dangerous NTC diagnostic findings directly: commanders who do not know they are being jammed because no one is watching the spectrum in real time. Col. Leslie Gorman, Army Capability Manager for EW, described the requirement as being able to “see yourself in the spectrum” — understanding what the unit’s own emission signature looks like to an adversary, and detecting when that signature is being exploited. Kenneth Strayer, product manager for EW and Cyber, extended the point: “Our systems also provide the ability to identify threats that are going to keep units from being able to communicate.” That sentence describes a fundamental shift from reactive to predictive spectrum management.


“Commanders will be able to see how they look in the spectrum and allow them to reposition or mask their own capabilities — or limit their EMS signature — while maintaining the ability to communicate.”
— Kenneth Strayer, PM EW & Cyber, PEO IEW&S

LAYER 2 — DETECTION: TLS BCT MANPACK, TEWS, AND ML SIGNAL CLASSIFICATION

Monitoring produces a picture. Detection produces an assessment. The TLS BCT Manpack — the Terrestrial Layer System Brigade Combat Team variant, now in rapid fielding to BCTs — provides the sensing capability that turns raw spectrum data into characterized threats. It combines signals intelligence collection, electronic attack, and direction finding in a manpack-portable system that enables brigades to identify adversary emitters, geolocate them for targeting, and begin electronic attack — all from a platform small enough to accompany a dismounted force.

What makes TLS BCT operationally relevant for the AI decision-support problem is its integration of machine learning signal recognition software. The system does not merely report that an unknown signal is present; it classifies the signal against known emitter libraries, assigns confidence levels to its attribution, and passes that structured intelligence into the NGC2 data layer where it can be correlated with other sensor feeds and acted upon. The Tactical Electronic Warfare System (TEWS), fielded to infantry BCT formations, provides the same ML-enabled signal recognition at the infantry echelon, ensuring that EW intelligence is not confined to armored or heavy formations.

At the more advanced end, DARPA’s PROWESS program — the Processor Reconfiguration for Wideband Sensor Systems, under development by SRI International and USC — is building reconfigurable processors that reclassify and respond to RF signals within 50 nanoseconds. This is not a human-speed capability; it is machine-speed autonomous threat characterization operating far faster than any human analyst can process. PROWESS is designed specifically for the problem that has broken traditional EW systems: modern digitally programmable radars and waveform-agile jammers change their signatures so rapidly that a library-lookup system built on known threats becomes obsolete almost as soon as it is deployed. AI-driven classification that learns from data rather than matching against fixed libraries is the only response at that speed.

The practical result for the operational commander is a detection picture that is continuous, automated, and increasingly accurate against novel threats — rather than dependent on a specialist analyst manually reviewing intercepts. At Project Convergence Capstone 5 in March 2025, the Army demonstrated that S2AS and TLS Manpack could feed adversary and friendly electromagnetic emissions data simultaneously into the NGC2 common data layer, providing a unified EMS picture that the division commander’s staff could see and act on. This was described by program officials as a proof of principle for the full operational vision.

System EMS Function AI/ML Component

• S2AS Real-time EMOE visualization, friendly signature monitoring, EMI detection Automated signal presence/anomaly alerts; EMI source attribution
• TLS BCT Manpack Brigade-level SIGINT collection, direction finding, EA ML signal classification against known/novel emitter libraries
• TEWS / TEWS-I Infantry BCT ES and limited EA ML signal recognition and IC-integrated signal detectors
• EWPMT-X / TAK-X Commander’s C2 tool for EMS planning, modeling, EA management ML-fused SIGINT; COA modeling; effects simulation; NG C2 integration
• EMBM-J DS (Palantir) Joint-level COA analysis, scheme of maneuver EMS risk scoring AI-automated joint EMS planning; COA scoring; wargaming engine
• PROWESS (DARPA) Autonomous RF signal classification and waveform adaptation Real-time ML reconfiguration at 50-nanosecond speed

LAYER 3 — DECISION SUPPORT: EWPMT-X, EMBM-J, AND AI-DRIVEN COA RECOMMENDATION

Monitoring and detection are necessary but not sufficient. The commander’s EMS decision-support gap is not primarily informational — it is cognitive. Even with a real-time spectrum picture and accurate threat attribution, the translation from “we are being jammed on this frequency by a probable Krasukha-class system at this grid” to “here is what you should do about it, in priority order, with tradeoffs” requires analytical work that currently sits with CEMA specialists. AI is the mechanism to automate that translation at machine speed.

EWPMT-X, the Army’s modernized electromagnetic warfare planning and management tool now migrating to the Tactical Assault Kit (TAK) framework for cross-service commonality, is where that analytical translation is embedded. Capability Drop 4 of the earlier EWPMT already added machine learning and AI tools that take raw SIGINT gathered across the battlefield and produce structured intelligence for EW cadres. EWPMT-X extends this substantially: it provides planning, modeling, and managing capabilities for multiple EW assets, simulates effects of potential electronic attack options, receives lines of bearing and sensor data from TLS and S2AS to produce EMOE visualizations, and — critically — synchronizes EW and spectrum management operations across intelligence, operations, and cyberspace for Multi-Domain Operations.

The EW Arsenal, a proof-of-concept knowledge repository demonstrated at Project Convergence Capstone 5, is designed as an adjunct to EWPMT-X that brings EW capabilities into the same conceptual framework that commanders use for kinetic fires. Bret Eddinger, Senior Engineer for Offensive EW at PEO IEW&S, explained the intent: “When we say ‘guns or bullets,’ the Army has a sense about what those things are. They know how to characterize the impacts of those munitions on targets. But what we need to realize is that electronic warfare is just another form of fire.” The EW Arsenal aims to give the commander and CEMA staff access to characterized EW effects against specific target types, rated by confidence and suppression duration, in the same mental model as artillery fire planning. An AI assistant that can propose “employ TEWS EA on this emitter to suppress for 15 minutes while you maneuver through this corridor” is a decision-support interface the combat arms commander can actually use without a CEMA translator.

At the joint level, the Palantir Electromagnetic Battle Management – Joint Decision Support (EMBM-J DS) prototype, awarded by DISA in March 2024 under a $9.8 million OTA agreement, represents the most explicitly AI-driven EMS decision-support tool in current development. Built on Palantir’s AIP (Artificial Intelligence Platform) at Impact Level 6, it ingests component-level courses of action and schemes of maneuver and evaluates the associated electromagnetic spectrum opportunities and risks against each. It produces COA scoring, supports wargaming of EMS effects, and automates key operational electromagnetic spectrum planning processes that currently require hours of manual staff work. EMBM-J DS Program Manager Betsy Park described it as “a first-of-its-kind planning tool that can automate key operational electromagnetic spectrum planning processes” — a description that, for an EMS-blind Army staff, is genuinely transformative.


“The ability to ingest component-level courses of action and schemes of maneuver into an overall joint plan and evaluate the associated electromagnetic spectrum opportunities and risks will be a significant technological leap forward for EMS operational planners.”
— Betsy Park, EMBM-J Program Manager, DISA PEO Spectrum

NGC2 AS THE INTEGRATING ARCHITECTURE

Each of the tools described above is powerful in isolation. What makes them operationally decisive is integration — and that is precisely what the Army’s Next Generation Command and Control (NGC2) initiative is designed to provide. NGC2 is not another system; it is a data-centric architecture for how all data, including EMS data, flows from sensor to decision-maker. Its design goal, as articulated by Kaloostian, director of the C2 cross-functional team within the new Transformation and Training Command, is to enable commanders to do “more, better, faster” in a DDIL environment where the old network architecture — heavy, slow to set up, and oblivious to spectrum signature management — would fail.

The NGC2 technology stack runs from a transport layer through an integration layer to a data layer to an application layer. It is the integration layer that is directly relevant to the EMS decision-support problem: this is where streams of information — S2AS detections, TLS intercepts, EWPMT-X planning data, intelligence feeds, friendly force positions — are fed into ML tools that triage and curate them before commanders interact with the data layer. The result, when working as designed, is a common operating picture in which EMS information is not a separate specialist domain but a visible layer in the same map interface the commander uses for maneuver, fires, and sustainment.

The Lightning Surge 1 exercise at Schofield Barracks in January 2026, conducted with the 25th Infantry Division, demonstrated the first operational proof of this integration. The Lockheed Martin-led NGC2 prototype — using a common data layer built by Raft with AI tools developed by Accelint — enabled the division to simultaneously visualize adversary and friendly electromagnetic emissions, share battlefield graphics and soldier positions, and process AI-generated spot reports using natural language voice and chat interfaces. The Army’s own after-action description identified “integrating and visualizing adversary and friendly electromagnetic emissions” and “accelerating commander updates using an AI pipeline to automate radio reporting” as two of the three primary assessed functions. EMS data was not a specialist annex; it was a first-class data stream in the division C2 picture.

The competing NGC2 prototype — Anduril’s development with the 4th Infantry Division, tested through the parallel Ivy Sting series at Fort Carson — provides a complementary picture. The Ivy Sting III exercise in December 2025 demonstrated AI models reviewing sensor data to rapidly recognize, process, and nominate targets for the fires kill chain. The application layer presents this to the commander as a series of decision applications covering intelligence, maneuver, fires, protection, sustainment, and information advantage, plus an operational modeling tool that generates courses of action using machine learning based on available data. The EMS picture flows into the same interface, alongside kinetic options, as one unified decision space.

THE DEGRADED ENVIRONMENT USE CASE: AI WHEN THE NETWORK IS FAILING

The most operationally demanding — and most important — use case for AI decision support is not when communications are working but when they are failing. The NTC repeatedly documents the catastrophic impact of commander cognitive paralysis when the common operating picture disappears under jamming. AI that can only function on a robust network is precisely wrong for this problem. The relevant question is what the system can do when bandwidth is degraded, nodes are jammed off, and the commander’s first-tier communications are denied.

NGC2’s architecture explicitly addresses this. At the NetModX experimentation event in September 2024, the Army tested the NGC2 architecture’s first experiment in a degraded environment — specifically examining how the system would function under a robust network, a resilient network, and an “intelligent, threat-informed network.” The concept of the threat-informed network is critical: an AI layer that monitors the spectrum, detects degradation, attributes it to probable cause (jamming versus interference versus equipment failure), and recommends network adaptation — routing around jammed nodes, switching waveforms, reducing emission signatures — automatically, without waiting for human analysis to complete.

L3Harris’s NGC2 contributions specifically address EW-contested communications continuity. Their architecture integrates multiple waveforms — TSM, Wraith, ARROW, MUOS, LDD, and FH3 — with AI-enabled adaptive gateway routing that automatically selects the available transport modality as the spectrum environment changes. The system provides waveform redundancy, dual-channel cross-banding, and iterative software-upgradeable countermeasures ensuring that the C2 network degrades gracefully rather than failing catastrophically when the first tier is jammed. The NGC2 Gateway Manpack, delivering up to 50 Mbps on a scalable mesh network with up to 800 nodes, is specifically designed to sustain mission command applications under disrupted conditions.

The S2AS contribution to degraded operations is equally direct. By detecting jamming at the moment it begins — not after communications have already failed — S2AS gives the commander a lead time window to act before cognitive paralysis sets in. If the AI system identifies a jamming signal consistent with Krasukha-class barrage jamming on the primary command net and immediately surfaces “Switch to PACE Contingency: SATCOM alternate; suppress emitter at [grid] with TEWS EA; reduce friendly CP emission profile” as a recommended action, the commander’s decision is reduced from open-ended analysis under stress to confirmation or rejection of a specific recommendation. That is the difference between paralysis and action.

The DDIL Scenario: What AI Decision Support Looks Like in Practice

0614 local. BCT command post detects sudden degradation on primary VHF command net. S2AS simultaneously alerts: new wideband emission detected, bearing 073 degrees, signal characteristics consistent with adversary EW system library entry KR-4 (Krasukha-class). Confidence: HIGH. TLS Manpack confirms direction finding fix. EWPMT-X automatically generates three recommended COAs ranked by mission impact: (1) Transition primary C2 to SATCOM alternate, suppress emitter with TEWS EA, maintain mission timeline — risk: exposes TEWS position; (2) Reduce CP emission profile, transition to burst-transmission protocol, request TLS EAB support from division — risk: 12-minute delay in fire mission; (3) Maneuver CP 800m to terrain mask, re-establish primary net from defilade — risk: 25-minute C2 gap. Commander selects COA 1. S2AS continues monitoring. Total elapsed time from jamming detection to commander decision: 90 seconds. Previous baseline without AI support: 8–15 minutes, if diagnosed correctly at all.

PACIFIC COMMAND’S AI SPECTRUM MANAGEMENT: THE INDOPACOM MODEL

Beyond the Army’s organic programs, U.S. Pacific Command (INDOPACOM) has already demonstrated what AI-enabled spectrum management looks like at the combatant command level — and the contrast with the prior baseline is stark. Bob Marcial, chief spectrum manager at INDOPACOM, described the old approach at TechNet Indo-Pacific 2025: “Right now, what we have are a lot of Excel spreadsheets, but AI is helping us do real-time spectrum management.” That sentence captures the before-and-after succinctly. The three AI-driven elements Marcial identified as transforming INDOPACOM’s spectrum operations are predictive pattern modeling of the EMS, real-time conflict detection and deconfliction, and waveform-adapting capabilities for interoperability with joint coalition partners.

Predictive pattern modeling is particularly relevant to the commander’s threat detection problem. An AI system trained on historical spectrum data from a given operational area can identify anomalous emission patterns — emissions that do not match known friendly or civilian signatures, that appear on frequencies associated with adversary EW doctrine, or that correlate temporally with known adversary operational activity cycles. This transforms threat detection from reactive (responding after jamming has begun) to anticipatory (identifying probable pre-jamming reconnaissance activity and alerting the commander before the attack).

THE DIGITAL ENEMY COMMANDER: AI RED-TEAMING THE EMS

A distinct but operationally powerful AI application sits at the intersection of threat detection and decision support: the use of AI agents to model adversary EW commander decision-making. The Army’s Military Intelligence Professional Bulletin published an analysis in mid-2025 describing this concept — AI agents trained on comprehensive datasets of adversary behavior, doctrine, EW communications, and decision-making patterns that can function as “digital enemy commanders.”

For EMS operations, this means an AI agent that models how a Russian EW battalion commander would respond to observed friendly force actions — what jamming assets would be activated, on what frequencies, at what point in the maneuver timeline, and with what intended effect. The intelligence officer can query this agent: “Given our planned Phase 2 movement through grid 47U, what EW response does the adversary 79th Separate Reconnaissance Brigade’s organic EW element most likely execute?” The agent’s response incorporates known Russian EW doctrine, observed signatures from prior contacts, and pattern analysis of adversary decision cycles.

Integrated with S2AS and EWPMT-X, this digital adversary modeling creates a closed prediction-confirmation loop: the AI predicts adversary EW actions based on friendly force activity; S2AS detects whether predicted emissions appear; EWPMT-X updates COA recommendations based on prediction accuracy. Over time, the model improves against the specific adversary unit the brigade is facing, not just against generic doctrine. This is the electronic warfare equivalent of what the 18th Airborne Corps’ AI-enabled targeting demonstrated for fires — dramatically increased precision and reduced cycle time against a specific, characterized threat.

THE RISKS: OVER-AUTOMATION, DATA POISONING, AND THE TRUST PROBLEM

The capabilities described above are real, are fielding, and represent a genuine transformation of the commander’s EMS decision problem. They also carry specific risks that any serious analysis must address.

The first risk is automation bias — the well-documented tendency of human decision-makers to over-trust AI recommendations, especially under stress. A commander who has been trained to act on EWPMT-X COA recommendations without scrutiny is vulnerable to a single point of failure: if the system’s threat classification is wrong — misidentifying friendly interference as adversary jamming, or mistaking a new commercial waveform for a threat emitter — the commander acts on bad intelligence at machine speed. Georgetown’s Center for Security and Emerging Technology flagged this specifically in its April 2025 analysis of military AI decision-support systems, noting that large language model-based systems can “confidently present incorrect information” and that skewed or scarce training data directly degrades battlefield recommendation quality.

The second risk is adversarial data manipulation. Russia is already using AI-based spectrum analysis tools against Ukrainian communications, including techniques designed to manipulate the electromagnetic environment in ways that exploit adversary classification systems. A sophisticated adversary who understands that the U.S. Army’s TLS BCT Manpack classifies signals against a known emitter library could, in principle, transmit spoofed signals calibrated to trigger specific AI classification outcomes — inducing the commander to take a predicted response that the adversary has planned for. This is a new vulnerability that did not exist when EW classification was done by human analysts applying judgment rather than by ML models applying pattern matching.

The third risk is the DDIL vulnerability of the AI system itself. An AI decision-support tool that requires cloud connectivity to function degrades in exactly the environment where it is most needed. The Army’s NGC2 design explicitly addresses this by emphasizing edge computing — ML tools that run on local compute within the unit’s data layer rather than requiring continuous cloud uplinks. But the residual capability of a disconnected AI decision-support system — what it can still do on local compute when the network is jammed and cloud sync is unavailable — is a critical parameter that must be defined, tested, and communicated to commanders before they encounter it in combat.


“The most transformative technology will have a bigger impact on future warfare than artificial intelligence. The goal is human decisions at machine speed.”
— Brig. Gen., Army NGC2 Program, December 2025

WHAT THE COMMANDER NEEDS TO DEMAND

Given this landscape, what should the operational commander demand from industry and program offices as AI-enabled EMS decision support matures? Three requirements stand out.

First, explainability. A COA recommendation that arrives without attribution — without a clear statement of what sensor data drove the threat classification, what confidence level the AI assigns to the emitter identification, and what doctrinal basis underlies the recommended response — is not actionable for a commander who needs to be able to override it. The EW Arsenal concept’s emphasis on characterizing EW effects the way artillery effects are characterized is the right model: the commander needs to understand what the AI is proposing to do and why, expressed in terms that do not require a CEMA specialist to interpret.

Second, graceful degradation specification. Every AI tool in this stack must have a documented and tested DDIL mode: what it does with partial sensor feeds, with degraded compute, with intermittent network connectivity. The commander must know in advance what the system can still do when 60% of its sensor inputs are offline — not discover it for the first time at 0300 during a contested river crossing.

Third, adversarial testing at realistic scale. TLS BCT Manpack, EWPMT-X, and EMBM-J DS have been tested in controlled exercises. They have not yet been tested against a sophisticated adversary running an active deception and spoofing campaign against the AI classification layer. The Army’s NGC2 2026 goal — testing the NGC2 prototype against enemy sensor capabilities — is the right next step. But it must include red teams specifically tasked with confusing and manipulating the AI components, not just stressing the network.

CONCLUSION: THE ARCHITECTURE IS READY; THE INTEGRATION IS THE WORK

The question Stephen asked — can AI systems help the operational commander monitor the electronic environment, detect threats, and recommend strategy and tactics to counter and operate in degraded conditions — has a clear and specific answer. Yes, and the tools are either in fielding or in FY2026 demonstration and procurement cycles. S2AS provides the real-time monitoring layer. TLS BCT Manpack and TEWS provide the ML-enabled detection and classification layer. EWPMT-X and the EW Arsenal provide the COA recommendation layer. EMBM-J DS provides the joint-level AI planning support. NGC2’s common data layer integrates all of them into a single picture that treats the electromagnetic battlespace as a first-class C2 domain rather than a specialist annex.

The gap is not technology — it is integration, tested DDIL performance, adversarial robustness, and the command culture change required to act on AI recommendations with the confidence and speed that closes the kill chain before the adversary’s 48-hour adaptation cycle makes the recommendation obsolete. Col. Gorman’s analogy is the right frame: owning the night took a generation of doctrine, training, and leader development after night vision goggles were invented. Owning the spectrum will require the same. The goggles are now being fielded. The work of building the Army that knows how to use them has barely begun.

SOURCES AND CITATIONS

S2AS — Spectrum Situational Awareness System

[1] PEO IEW&S. "Transformation in Contact Units Test Spectrum Situational Awareness System." August 20, 2025. https://peoiews.army.mil/2025/08/20/transformation-in-contact-units-test-spectrum-situational-awareness-system/

[2] Lawrence, Drew F. "Army Awards Prototype for Spectrum Sensing and Visualization Tool." DefenseScoop, April 2, 2025. https://defensescoop.com/2025/04/02/army-s2as-3db-labs-award-prototype-spectrum-awareness-system/

[3] Lawrence, Drew F. "Army Spectrum Tool Will Feature in Upcoming TiC Exercises to Inform Rapid Fielding Decision." DefenseScoop, February 24, 2025. https://defensescoop.com/2025/02/24/army-spectrum-tool-transforming-in-contact-rapid-fielding/

[4] Janes. "US Army to Field Key EW Programme by 2026." December 17, 2024. https://www.janes.com/osint-insights/defence-news/c4isr/us-army-to-field-key-ew-programme-by-2026

TLS, TEWS, EWPMT-X — EW Sensing and Planning Tools

[5] PEO IEW&S PM EW&C. Program Fact Sheet, 2025. https://peoiews.army.mil/pm-ewc/

[6] PEO IEW&S. "Fusing Intel and EW Data into the Army’s Data Centric NGC2 Architecture." December 23, 2024. https://peoiews.army.mil/2024/12/23/278296/

[7] PEO IEW&S. "PEO IEW&S Showcases the Future of Intelligence, EW and Surveillance at PC-C5." April 25, 2025. https://peoiews.army.mil/2025/04/25/peo-iews-showcases-the-future-of-intelligence-electromagnetic-warfare-and-surveillance-at-pc-c5/

[8] Lawrence, Drew F. "Army Alters Approach for Electromagnetic Spectrum Planning Tool." DefenseScoop, April 30, 2024. https://defensescoop.com/2024/04/30/army-alters-approach-electromagnetic-spectrum-planning-tool/

[9] Armada International. "Spectrum Battle Management." December 4, 2023. https://www.armadainternational.com/2023/12/us-army-ewpmt-moves-forward/

[10] Gorman, Col. Leslie. "Owning the Spectrum: How the Army Is Reinventing Electromagnetic Warfare." FedGov Today, September 11, 2025. https://fedgovtoday.com/fedgov-blogs/owning-the-spectrum-how-the-army-is-reinventing-electromagnetic-warfare

Palantir EMBM-J DS

[11] Palantir Technologies. "Palantir to Deliver Electromagnetic Battle Management – Joint Decision Support (EMBM-J DS) Prototype to DISA." March 29, 2024. https://investors.palantir.com/news-details/2024/Palantir-to-Deliver-Electromagnetic-Battle-Management---Joint-Decision-Support-EMBM-J-DS-Prototype-to-the-Defense-Information-Systems-Agency-DISA/

[12] Military Embedded Systems. "Electromagnetic Battle Management Prototype to Be Developed for DISA by Palantir." 2024. https://militaryembedded.com/radar-ew/rf-and-microwave/electromagnetic-battle-management-prototype-to-be-developed-for-disa-by-palantir

[13] The Defense Post. "DISA Orders Electromagnetic Battle Management Prototype From Palantir." April 4, 2024. https://thedefensepost.com/2024/04/03/us-electromagnetic-battle-management-prototype/

NGC2 — Next Generation Command and Control

[14] DefenseScoop. "How the Army Built Next-Gen Command and Control." March 20, 2025. https://defensescoop.com/2025/03/20/how-army-built-next-gen-command-and-control-ngc2/

[15] Federal News Network. "Inside the Army’s Push to Scale Next Gen C2 System and Prepare for AI." October 31, 2025. https://federalnewsnetwork.com/federal-insights/2025/10/inside-the-armys-push-to-scale-next-gen-c2-system-and-prepare-for-ai/

[16] Breaking Defense. "What’s Next for Army’s Ambitious Next Gen C2 Effort: 2026 Preview." December 31, 2025. https://breakingdefense.com/2025/12/whats-next-for-armys-ambitious-next-gen-c2-effort-2026-preview/

[17] Breaking Defense. "Army Tests Next-Gen C2 Data Layer for the First Time." January 23, 2026. https://breakingdefense.com/2026/01/army-tests-next-gen-c2-data-layer-for-the-first-time/

[18] Lockheed Martin. "Lockheed Martin Accelerates NGC2 Capabilities at Lightning Surge 1." January 23, 2026. https://news.lockheedmartin.com/2026-01-23-Lockheed-Martin-Accelerates-Next-Generation-Command-and-Control-NGC2-Capabilities-and-Real-Time-Decision-Making-at-Lightning-Surge-1

[19] Army.mil. "Army Teams with Industry to Refine AI Potential Supporting C2." December 14, 2025. https://www.army.mil/article/289556/army_teams_with_industry_to_refine_ai_potential_supporting_command_and_control

[20] DefenseScoop. "Data and Integration Will Be Core of Army’s Next-Gen C2." June 6, 2025. https://defensescoop.com/2025/06/06/data-and-integration-will-be-core-of-armys-next-gen-c2/

[21] L3Harris. "NGC2: Enabling Next Generation Command and Control." 2025. https://www.l3harris.com/all-capabilities/ngc2-enabling-next-generation-command-and-control

AI Decision Support: Risks and Frameworks

[22] Probasco, Emelia et al. "AI for Military Decision-Making: Harnessing the Advantages and Avoiding the Risks." CSET, April 2025. https://cset.georgetown.edu/publication/ai-for-military-decision-making/

[23] Williams, LTC Timothy J. "AI-Enabled Wargaming at the U.S. Army Command and General Staff College." Small Wars Journal, January 16, 2026. https://smallwarsjournal.com/2026/01/16/ai-enabled-wargaming-cgsc/

Cognitive EW and INDOPACOM Spectrum Management

[24] AFCEA SIGNAL. "Artificial Intelligence-Driven Components Helping Advance EMS Space." TechNet Indo-Pacific 2025. https://www.afcea.org/signal-media/artificial-intelligence-driven-components-helping-advance-ems-space

[25] USNI Proceedings. Lt. Cmdr. Brian P. Gannon. "Implement AI in Electromagnetic Spectrum Operations." August 2023. https://www.usni.org/magazines/proceedings/2023/august/implement-ai-electromagnetic-spectrum-operations

Digital Enemy Commander / AI Red Team

[26] Military Intelligence Professional Bulletin. "Know Thy Enemy: Using AI to Create Digital Enemy Commanders." July–December 2025. https://mipb.ikn.army.mil/issues/jul-dec-2025/know-thy-enemy/

PROWESS / Cognitive EW Research

[27] Air Force Tech Connect. "Artificial Intelligence and Machine Learning Aim to Boost Tempo of Military Operations." https://airforcetechconnect.org/news/artificial-intelligence-and-machine-learning-aim-boost-tempo-military-operations


SIGNAL Magazine is published by AFCEA International • 4400 Fair Lakes Court • Fairfax, Virginia 22033 • www.afcea.org