AI / ML | Soldier Systems Daily Soldier Systems Daily

Archive for the ‘AI / ML’ Category

Accrete Expands Classified Data Support to Meet Growing Customer Demand and Mission Needs

Wednesday, February 25th, 2026

NEW YORK, Feb. 23, 2026 — Accrete, Inc., a dual-use AI company founded in 2017, whose Knowledge Engine platform encodes institutional knowledge, judgment, and expertise into AI agents for decision superiority, announced that it has successfully deployed its Argus platform into the Joint Worldwide Intelligence Communications System (JWICS) through Second Front’s 2F Game Warden®.

This is an important step in Accrete’s ongoing effort to meet surging demand by delivering expanded capabilities for teams operating in Top Secret data environments. This achievement reinforces Accrete’s dedication to current and future defense stakeholders while bringing more value to organizations, reducing friction between workflows, and accelerating the delivery of insights that inform time-sensitive decisions.

“Customer need for Accrete’s platform is growing because mission teams need trusted, actionable answers from complex data,” said Prashant Bhuyan, Founder, CEO, and Chairman of Accrete, Inc. “JWICS deployment means that Accrete’s Expert AI Agents can interact with open source and government data to drive faster, better-informed outcomes, delivering decision advantage in mission-critical environments.”

Accrete’s Knowledge Engine is the foundational cognitive infrastructure for encoding institutional knowledge into persistent organizational intelligence. Knowledge Engines become even more powerful when they can operate across classified and unclassified data, connecting siloed systems into shared context while preserving the why behind past decisions. Accrete’s platform enables trusted AI agents to surface critical insights and emerging risks that humans would never think to search for as information complexity grows.

“This milestone reflects Accrete’s broader focus on scaling alongside customer needs. The company has been operational at DoW Impact Level 4 (IL4) and is now expanding secure support for classified data, accelerating capability delivery, and ensuring defense customers can apply Accrete’s AI platform across the full range of mission operations,” said Bill Wall, CEO of Accrete AI Government.

Posted in AI / ML, Press Release | No Comments »

Army Doctrine Writers Embrace AI to Speed Knowledge to the Force

Wednesday, February 25th, 2026

FORT LEAVENWORTH, Kan. – The process of writing Army doctrine, traditionally measured in years, is getting a 21st-century upgrade. Thanks to digital tools, some enabled by artificial intelligence, authors at the Combined Arms Doctrine Directorate are examining processes and gaining efficiencies without sacrificing their high standards for quality and precision.

“We have had people ask us about using AI and large language models to speed up the doctrine development process for years,” said CADD Director Richard Creed, Jr. “So, when some of these tools became available the first thing we did was figure out their capabilities. That meant that we needed to identify people on our team with some AI experience from using civilian applications who could help us figure out what was possible.”

Two officers, Lt. Col. Scott McMahan, a doctrine writer in the Operational Level Doctrine Division, and Maj. Matthew Martinez, CADD’s Information Management Officer, took leaders’ guidance and developed a four-pronged strategy to equip every doctrine writer with the skills and tools necessary to leverage AI effectively. The strategy began with foundational training for all CADD members, allowing them to apply approved AI tools to their work immediately.

To foster expertise, the plan also calls for training a “master gunner,” someone highly qualified in using AI tools, within each doctrine division. These experts can then help their teams tackle more difficult challenges and find advanced applications for the technology. Furthering the integration, AI best practices are being formally incorporated into the Doctrine Developer’s Course, ensuring future writers are introduced to these techniques from the start.

Finally, CADD leaders are working with the Combined Arms Command and industry partners to build a purpose-made AI tool. This new software will be designed for all Army doctrine writers, but it will be especially helpful to authors at the Centers of Excellence, where personnel and time are often scarce.

According to McMahan, the answer is not in “letting AI write the books” that guide the force. Instead, it is about providing human experts with powerful new assistants. “You treat it like a resourceful and motivated young officer who might not know all the information, but they can certainly assist you in cutting some corners and being a little more efficient,” he said. “The bottom line is that eventually it should, even if you’re just shaving at the margins, it’s going to speed up the time when doctrine will reach Soldiers in the field.”

Creed said his guidance was always to treat technology as a tool, not a panacea. “Because we have such good people, I expected that we would stay abreast of developments and be ready when the tools were ready,” he said. “They came up with an approach that would make one person in each doctrine division an AI subject matter expert while training everyone else in the basic use of AI for tasks related to their job. It was no different than ensuring all Soldiers are combat lifesavers while assigning an actual medic to each platoon.”

So far, changes have been small, but those small changes add up when looking at multiple individual and project tasks. One example is an internally developed tool that gives writers the ability to quickly search in hundreds of texts for historical vignettes that illustrate a complex doctrinal point. This task could once have taken days of research, especially for a new doctrine author.

“The large language model tools under development now have access to the databases we needed access to in the past. Access to the data is the foundational measure of whether the tools are useful to us. We tip our hats to those who figured out how to do that,” said Creed.

Another use of AI is something McMahan calls “breaking the blank page.” The tool can help writers get through creative blocks and generate ideas. “We were looking for some more meat for an idea,” he said. “We were able to feed this tool some initial thoughts, and of the three paragraphs it spit out, one sentence was used, but that was a really powerful and useful sentence.”

Blending technology and in-house experience is reducing administrative burdens for doctrine authors and staff in several ways. For instance, digital tools that assist with grammar and readability free authors to focus on the complex aspects of their work, which in turn saves valuable time for CADD’s high-demand editors. This efficiency extends beyond writing, as leaders have improved the publications tracking process with automated forms and use of Power Business Intelligence, and individuals are creating their own self-study tools to prepare for internal certification.

McMahan recognizes the technology is not perfect, but notes that it is improving over time. He said that AI models can “hallucinate” by inventing facts or confuse source materials when asked a question. These are critical flaws in a field where accuracy is paramount. In one case, an AI-generated question for a doctrine test was based on an outdated manual, an error that was only caught because the user creating the test was an expert on the topic.

Situations like this drive home why subject matter expertise is paramount. “We made it perfectly clear that AI tools were not intended to be a crutch for not doing the work we expect from our people,” Creed said. “Humans will review every line of what an LLM produces for accuracy. To make sure that happens one must make sure your people know their business.”

To learn more about Army doctrine and the Combined Arms Doctrine Directorate visit the Combined Arms Command’s website.

By Randi Stenson, MCCoE Public Affairs

Posted in AI / ML, Army, Guest Post | No Comments »

Pennsylvania Guard Soldiers Strengthen AI, Critical Thinking Skills

Sunday, February 22nd, 2026

FORT INDIANTOWN GAP, Pa. — Pennsylvania National Guard Soldiers and civilian employees participated in an Artificial Intelligence 201 course Feb. 11–12.

The course, taught by U.S. Army War College faculty, aimed to prepare leaders to responsibly integrate artificial intelligence into military decision-making while reinforcing critical thinking and mission command principles.

“A lot of people find it very scary, and just like with any new technology, we should be cautious,” said Lt. Col. Kelly Ihme, an assistant professor at the U.S. Army War College and one of the course’s instructors. “Trust but verify. But this is a computer program. It’s predictive math.”

AI 101 introduces Soldiers to the basics of AI, including how to use and understand it, while AI 201 focuses on critical thinking and effective AI prompting.

“At the War College, we’re starting to develop that type of course,” Ihme said. “It’s more about, how do we think about problems? And then where does AI get inserted into those problem sets? So that we’re never taking the human out of the loop, but we’re stepping up the critical thinking and really engaging critical skills questions on AI and where it fits.”

The students found the classes useful and knowledgeable, even wanting to attend more in the future.

“If they had another AI class, I would go again,” said Maj. Maria Myers, logistics branch chief at Joint Force Headquarters. “I would even do this exact same level again, just for repetition and to make sure that I’m still using it correctly, because I do plan to try to implement it in a couple of ways already that they had mentioned.”

Ihme said the 201 course pushes students to dig deeper than the nuts and bolts of AI. The class encourages them to approach AI with a leadership mindset.

“It’s a mission command skill. And if you’re not using AI with some of those mission command ideals in mind, you’re going to get icky outputs or less efficient outputs, and then you’re going to put AI aside,” said Ihme. “You’re not going to throw a private aside and tell them they’re an ineffective private because you didn’t provide them the leadership they needed to succeed. It’s the same with AI.”

Myers said that while she believes AI should be used with caution, it remains a helpful tool for saving time. But, she emphasized that users need to understand how to properly use and prompt AI before utilizing it for work.

“I’m sure some people will think it’s a pain, and there will be people that are against it, because it’s different,” Myers said. “But I think that this class and the level it goes into is a good introductory class because you need to have the knowledge and the understanding and the information from the human aspect to get the appropriate answer.”

The Pennsylvania National Guard is hoping to expand these classes in an effort to have Fort Indiantown Gap become an AI center of excellence, Imhe said. Ihme said she is ready for the next step of AI in the Army.

“Let’s keep providing classes like this, both foundational and thinking classes, so that we can figure out, not just that we’re playing with AI, but that we’re taking the next leap of using AI to propel us forward,” Ihme said.

By SGT Kayden Bedwell

Posted in AI / ML, Army, Guest Post | No Comments »

ASI Awarded Defense Innovation Unit (DIU) Prototype OTA for AI-Enabled Joint Sustainment Decision Tool (JSDT)

Friday, February 20th, 2026

BOSTON, Feb. 19, 2026 — Air Space Intelligence (ASI), a pioneer in AI-powered decision support software for the world’s most demanding domains, today announced it has been awarded a contract by the Defense Innovation Unit (DIU) to configure and implement advanced logistics decision-support for leaders operating in complex, multi-domain operational environments.

Under the award, ASI will deliver its commercial software capability for decision support, capable of generating realistic and actionable courses of action (COAs) for decision makers facing operational uncertainty and adversary pressure. ASI’s solution will rapidly analyze evolving conditions, identify critical dependencies across the logistics kill chain, and present commanders with viable options that account for disruption, attrition, and constrained resources.

Modern military operations are increasingly shaped by the ability, or inability, to sustain forces. Adversaries actively target logistics networks, infrastructure, and decision timelines, turning sustainment into a decisive element of the fight. ASI’s solution reflects this reality by explicitly modeling logistics as an operational system, enabling commanders to understand how logistics decisions impact operational endurance, tempo, and risk over time.

“Commanders today must make decisions faster than ever, often with incomplete information and under persistent threat to the status, suitability and capacity of their nodes and distribution networks.,” said Mark Lepczyk, President of ASI Federal. “This award from DIU recognizes the need for AI-centric solutions that don’t just optimize a single plan, but help leaders explore multiple courses of action and understand how each one stresses or preserves the logistics kill chain.”

ASI’s commercial solution is designed to support decision-making in dynamic environments where traditional assumptions break down—where resupply is uncertain, infrastructure is contested, and degradation is the norm rather than the exception. By combining AI-driven analysis with domain-informed models, the solution helps commanders evaluate tradeoffs between speed, risk, sustainability, and mission outcomes.

This DIU award reflects the Department of War’s continued focus on becoming an “AI-First” warfighting force. ASI’s work directly supports this mission by applying advanced analytics and machine learning to command-and-control and sustainment problems that increasingly determine operational success.

Posted in AI / ML, Contracts, Press Release | Comments Off on ASI Awarded Defense Innovation Unit (DIU) Prototype OTA for AI-Enabled Joint Sustainment Decision Tool (JSDT)

Raft Awarded Contract for Desert Sentry to Accelerate Operator-Centric AI for CENTCOM Missions

Saturday, January 24th, 2026

Raft’s AI Mission System empowers CENTCOM operators with on-demand computer vision tailored to real-world missions – broad-area satellite search, distributed maritime monitoring, and counter-UxS threat detection.

MCLEAN, Va., Jan. 22, 2026 — Raft, a leading defense technology company, powering the software-driven foundation of modern warfare, has been awarded a highly competitive Other Transaction Agreement (OTA) by the Chief Digital and Artificial Intelligence Office (CDAO), in partnership with U.S. Central Command (CENTCOM), following a five-day vendor bake-off. Hosted last fall by CENTCOM, the evaluation sought cutting-edge solutions to operationalize AI on the battlefield. Raft’s win underscores its ability to rapidly deliver deployable, operator-ready technology that meets the evolving demands of the modern fight.

Raft’s AI Mission System ([R] AIMS) delivers a no-code machine learning system that puts the power of computer vision directly into the hands of operators. Raft built and deployed a fully containerized, Agentic AI platform that enables operators to train, evaluate, and deploy new Computer Vision models with confidence and verifiability—no data science background required.

“This wasn’t about building another tool,” said Shubhi Mishra, founder and CEO of Raft. “This was about rethinking how AI gets built for mission-critical environments and how we empower operators to adapt when the mission demands it. Raft exists to turn operators into super-operators, and that starts by putting AI creation directly in their hands.”

By integrating seamlessly with NGA Maven workflows, [R]AIMS enables model refresh in days—supporting key use cases like:

Satellite Broad Area Search: Quickly train gap-filler models on-the-fly to support large-scale overhead imagery tasks

Distributed Monitoring: Enable task forces like TF-59 to deploy and adapt CV models across maritime environments

Counter-UxS Threat Detection: Build computer vision to detect, label, and characterize optical threats without waiting on centralized updates

Built with Raft’s agentic AI principles at its core, the system features built-in guardrails via [R]AIMS and machine-assisted feedback loops for Responsible AI—ensuring every model is traceable, adaptable, and aligned with mission intent.

“[R]AIMS Vision enables something we’ve been chasing for years: COCOMs building their own AI at the speed of the mission,” said Bhaarat Sharma, CTO at Raft. “This is how you scale intelligence across the force—by putting tools in the hands of operators and giving them the autonomy to iterate in real time.”

[R]AIMS validates what is possible when Agentic AI moves beyond conversational interfaces to deliver autonomous, operator-controlled computer vision that adapts at mission speed. Operators rapidly created new gap-filler models for unplanned ISR scenarios—without needing data scientists on call. The system even handled optical threat detection and object tracking in real time, including NITF and other complex formats. [R]AIMS is now positioned for rapid expansion across combatant commands, supporting faster intelligence cycles, better decision-making, and real-time adaptation on the battlefield.

For more information about Raft, visit www.teamraft.com.

Posted in AI / ML, Industry, Press Release | Comments Off on Raft Awarded Contract for Desert Sentry to Accelerate Operator-Centric AI for CENTCOM Missions

JMRC Trains World-Class OC/Ts

Tuesday, January 13th, 2026

HOHENFELS TRAINING AREA, Germany – The Joint Multinational Readiness Center (JMRC) has been training its Observer, Controller/Trainers (OC/T) in new technologies to keep its world-class training ready for the future fight.

The OC/Ts from the nine “critter” teams are being trained in evolving tasks and technologies such as Maven, electronic warfare (EW), unmanned aircraft systems (UAS), and integrated tactical network (ITN).

“This OC/T recertification training is an opportunity for the OC/Ts to become proficient on some tasks that are constantly changing and new,” said Maj. Dustin Allen, deputy operations for JMRC. “It’s to meet higher headquarters’ intents of knowing new technologies so that we can better facilitate the rotational units as they come through ‘the Box’.”

One of the systems that the OC/Ts are training on is the Maven Smart System. The Maven Smart System is the Department of Defense’s most prominent artificial intelligence capability. Designed to process drone imagery and full-motion video, Maven integrates sensors with artificial intelligence and machine learning to enhance battlefield awareness and support operations such as targeting, logistics planning and predicting supply requirements for deployed Soldiers.

“Maven is something that is near and dear to (U.S. Army Europe and Africa’s) heart,” said Allen. “We are trying to incorporate it into our daily battle rhythms, so that the critter teams are well versed in it. We also want to be able to teach the rotational units that come through that aren’t familiar with the system and get them better with it.”

OC/Ts have also been working with UAS. UAS training develops Soldiers’ abilities to operate and employ aerial systems in support of reconnaissance, intelligence collection, and mission planning, helping to facilitate the combined arms fight on the ground.

“UAS is a big push, especially in past rotations where we have seen a massive increase in UAS capabilities on the battlefield,” said Allen. “OC/Ts are going to have their own UAS so that they can send a drone up and inject it there, and watch the rotational units’ UAS.”

Another system that has been seen on the battlefield that OC/Ts are being trained on is EW. EW enhances commanders’ abilities to detect, disrupt and protect against enemy electromagnetic capabilities, enabling freedom-of-action across the battlefield.

“Big in current warfare is the introduction of electronic warfare,” said Allen. “During this time, we are giving the OC/Ts the opportunity to see and become familiar with the vastly growing EW capabilities.”

The last system that the OC/Ts are being trained on is the ITN. ITN delivers secure, resilient and expeditionary communications that connect Soldiers, platforms and command posts across the battlefield. ITN is designed to operate in contested and degraded environments, and enable timely data sharing and mission command to support multi-domain operations.

“You can interconnect the radios that we use for our communications network across Hohenfels and JMRC, so we can communicate more clearly across the box,” said Allen.

“Our OC/Ts are already world-class,” said Allen. “This training can help make them even better and have more systems that they are proficient in. We’re really going to see all this hard work they are putting in, be used in the next Combine Resolve we host, and I’m excited to see them use all these new systems we have.”

Story by SGT Collin Mackall

7th Army Training Command

Posted in AI / ML, Army, Drone, EW, Guest Post, Training | Comments Off on JMRC Trains World-Class OC/Ts

Magnet Defense Enters into Definitive Agreement to Acquire ATG to Accelerate Integration of AI-Enabled Autonomy Solutions for National Security

Monday, January 12th, 2026

MIAMI, Jan. 9, 2026 — Today, Magnet Defense LLC, a developer of fully autonomous national security maritime platforms for fleet operations and missile defense missions, announces that it has officially entered into a definitive agreement to acquire Advanced Technology Group (ATG), subject to customary closing conditions. This acquisition further enhances Magnet Defense’s autonomy solutions by incorporating ATG’s open-architecture AI solutions into its DefendAI battlespace management suite. These are the brains and backbone behind Magnet Defense’s end-to-end AI-enabled autonomous maritime defense solutions.

ATG delivers end-to-end development and integration of advanced Command and Control and Artificial Intelligence capabilities for air, space, maritime, and surface platforms. With an elite group of mission architects, AI integrators, and software engineers, ATG is solving some of the U.S. Department of War’s most difficult challenges across all domains. ATG’s capabilities will accelerate Magnet Defense’s seamless integration of its platforms into theater and operational battlespace management systems.

Magnet Defense intends to begin aligning and integrating ATG’s Autonomy Exchange for Interoperable Modularity (AXIOM) AI-enabled autonomy stack with its own proven autonomy capabilities. AXIOM’s set of proven mission modules will streamline Magnet Defense’s integration into the native command and control systems found in military services, operations centers, and combatant commands across the sea, land, air, space, and cyber domains. ATG’s leadership and employees will continue to support existing customers while contributing to expanded programs across the combined organization.

About Magnet Defense
Magnet Defense is a developer of fully autonomous national security maritime platforms for fleet operations and missile defense missions. We integrate AI-driven software solutions, advanced manufacturing systems, and mission architecture expertise to deliver the most advanced purpose-built USVs for the U.S. and allied militaries. Learn more at www.magnetdefense.com

Posted in AI / ML, Industry, Maritime, Press Release | Comments Off on Magnet Defense Enters into Definitive Agreement to Acquire ATG to Accelerate Integration of AI-Enabled Autonomy Solutions for National Security

AI in Battle Management: A Collaborative Effort Across Borders

Thursday, January 8th, 2026

The 2025 series of the Decision Advantage Sprint for Human-Machine Teaming marked a significant step forward in the integration of artificial intelligence and machine learning into battle management operations. Through a series of groundbreaking experiments, including the recent DASH 3 iteration, the U.S. Air Force, alongside its coalition partners, Canada and the United Kingdom, tested and refined AI’s potential to enhance decision-making, improve operational efficiency, and strengthen interoperability in the face of growing global security challenges.

Held at the unclassified location of the Shadow Operations Center-Nellis in downtown Las Vegas, DASH 3 set the stage for this collaboration, led by the Advanced Battle Management System Cross-Functional Team. The experiment was executed in partnership with the Air Force Research Lab’s 711th Human Performance Wing, U.S. Space Force, and the 805th Combat Training Squadron, also known as the ShOC-N, further solidifying the commitment to advancing battle management capabilities for the future.

AI Integration into Operational Decision-Making

In the third iteration of the DASH series seven teams, six from industry teams and one from the ShOC-N innovation team partnered with U.S., Canadian, and U.K. operators to test a range of decision advantage tools aimed at enhancing the rapid and effective generation of battle course of actions with multiple paths. The goal of a Battle COA is to map sequences of actions that align with the commander’s intent while overcoming the complexities of modern warfare, including the fog and friction of battle. Examples of Battle COAs include recommended solutions for long-range kill chains, electromagnetic battle management problems, space and cyber challenges, or agile combat employment such as re-basing aircraft.

U.S. Air Force Col. John Ohlund, ABMS Cross Functional Team lead overseeing capability development, explained the importance of flexibility in COA generation: “For example, a bomber may be able to attack from multiple avenues of approach, each presenting unique risks and requires different supporting assets such as cyber, ISR [intelligence, surveillance, and reconnaissance], refueling, and air defense suppression. Machines can generate multiple paths, supporting assets, compounding uncertainties, timing, and more. Machines provide a rich solution space where many COAs are explored, but only some are executed, ensuring options remain open as the situation develops.”

This ability to explore multiple COAs simultaneously allows for faster adaptation to unforeseen challenges and provides operators with diverse strategies to act upon as the situation unfolds. AI’s integration into this process aims to not only speed up the decision-making cycle but also increase the quality of the solutions generated.

AI Speeds Decision Advantage

The speed at which AI systems can generate actionable recommendations is proving to be a game-changer in the decision-making process. Transitioning from the manual creation of COAs that once took minutes or tens of minutes to producing viable options in just tens of seconds was identified as a radical advantage in combat scenarios. Initial results from the DASH 3 experiment show the power of AI in enabling faster, more efficient decision-making.

“AI systems demonstrated the ability to generate multi-domain COAs considering risk, fuel, time constraints, force packaging, and geospatial routing in under one minute,” said Ohlund. “These machine-generated recommendations were up to 90% faster than traditional methods, with the best in machine-class solutions showing 97% viability and tactical validity.”

For comparison, human performance in generating courses of action typically took around 19 minutes, with only 48% of the options being considered viable and tactically valid.

“This dramatic reduction in time and improvement in the quality of solutions underscores AI’s potential to significantly enhance the speed and accuracy of the decision-making process, while still allowing humans to make the final decisions on the battlefield,” Ohlund added.

The ability to quickly generate multiple viable COAs not only improves the speed of decision-making but also gives commanders more options to work within a compressed time frame, making AI an essential tool for maintaining a strategic advantage in fast-paced combat situations.

Building Trust in AI: From Skepticism to Confidence

Skepticism surrounding the integration of AI in operational decision-making was common at the start of the DASH 3 experiment. However, participating operators saw a notable shift in their perspectives as the DASH progressed. U.S. Air Force First Lt. Ashley Nguyen, 964th Airborne Air Control Squadron DASH 3 participant, expressed initial doubt about the role AI could play in such a complex process. “I was skeptical about technology being integrated into decision-making, given how difficult and nuanced battle COA building can be,” said Nguyen. “But working with the tools, I saw how user-friendly and timesaving they could be. The AI didn’t replace us; it gave us a solid starting point to build from.”

As the experiment unfolded, trust in AI steadily increased. Operators, gaining more hands-on experience, began to see the value in the AI’s ability to generate viable solutions at an unprecedented speed. “Some of the AI-generated outputs were about 80% solutions,” said Nguyen. “They weren’t perfect, but they were a good foundation. This increased my trust in the system; AI became a helpful tool in generating a starting point for decision-making.”

Trust and Collaboration Across Nations

The collaboration between the U.S. and its coalition partners was highlighted throughout the 2025 DASH series. The inclusion of operators from the UK and Canada brought invaluable perspectives, ensuring that the decision support tools tested could address a broad range of operational requirements.

“We understand that the next conflict cannot be won alone without the help of machine teammates and supported by our allies,” said Royal Canadian Air Force Capt. Dennis Williams, RCAF DASH 3 participant. “DASH 3 demonstrated the value of these partnerships as we worked together in a coalition-led, simulated combat scenario. The tools we tested are vital for maintaining a decision advantage, and we look forward to expanding this collaboration in future DASH events.”

This integration of human-machine teaming and coalition participation highlighted the potential for improving multinational interoperability in the command-and-control battlespace. “The involvement of our coalition partners was crucial, not just for the success of DASH 3 but also for reinforcing the alliances that underpin global security. DASH experimentation is intentionally a low barrier for entry from a security classification standpoint, enabling broad participation from allies and coalition partners alike,” said U.S. Air Force Lt. Col. Shawn Finney, commander of the 805th Combat Training Squadron/ShOC-N.

Addressing Challenges: Weather and AI Hallucinations

The DASH 3 experiment was not just a test of new AI tools, but a continuation of a concerted effort to tackle persistent challenges, including the integration of weather data and the potential for AI “hallucinations.” These issues have been focus areas throughout the DASH series, with each iteration bringing new insights and refinements to ensure AI systems are operationally effective.

Weather-related challenges are a critical factor in real-world operations, but due to simulation limitations, they were not fully integrated in the DASH series. Instead, weather-related challenges were manually simulated by human operators through ‘white carding’, a method that provided scenario-based weather effects, such as airfield closures or delays, into the experiment.

“We didn’t overlook the role of weather,” explained Ohlund. “While it wasn’t a primary focus of this experiment, we fully understand its operational impact and are committed to integrating weather data into future decision-making models.”

The risk of AI hallucinations, instances where AI produces incorrect or irrelevant outputs, particularly when using large language models, was another challenge tackled during the DASH 3 experiment. Aware of this potential issue, the development teams took proactive steps to design AI tools that minimized the risk of hallucinations and organizers diligently monitored the outputs throughout the experiment.

“Our team didn’t observe hallucinations during the experiment, underscoring the effectiveness of the AI systems employed during the experiment,” said Ohlund. “While this is a positive outcome, we remain vigilant about the potential risks, particularly when utilizing LLMs that may not be trained on military-specific jargon and acronyms. We are actively refining our systems to mitigate these risks and ensure AI outputs are reliable and relevant.”

Looking Ahead: Building Trust in AI for Future Operations

As the U.S. Air Force moves forward with the 2026 series of DASH experiments, the lessons learned from 2025 iterations will serve as a crucial foundation for future efforts. The growing trust in human-machine collaboration, the strengthening of international partnerships, and the continuous refinement of AI tools all point to a future where AI plays an integral role in operational decision-making.

“The 2025 DASH series has established a strong foundation for future experiments, with the potential to further expand AI’s role in battle management,” said Ohlund. “By continuing to build trust with operators, improve AI systems, and foster international cooperation, the U.S. and its allies are taking critical steps toward ensuring they are prepared to address the evolving challenges of modern warfare.”

“This is just the beginning,” said Williams. “The more we can integrate AI into the decision-making process, the more time we can free up to focus on the human aspects of warfare. These tools are key to staying ahead of our adversaries and maintaining peace and stability on a global scale.”

Deb Henley

505th Command and Control Wing

Public Affairs

Posted in AI / ML, Air Force, C5ISR, Guest Post | Comments Off on AI in Battle Management: A Collaborative Effort Across Borders