DefenseNews: In a landmark experiment that could reshape how future wars are planned and fought, the U.S. Air Force announced that artificial intelligence (AI) tools outperformed human military planners in a major battle management experiment.
The results of the test part of the Air Force’s ongoing push toward human-machine teaming and AI-enhanced command and control systems show that AI can not only generate planning solutions faster than experienced officers but can also produce higher-quality tactical options with fewer errors.
The experiment was conducted under the Air Force’s Decision Advantage Sprint for Human-Machine Teaming (DASH) program, a series of wargame-like trials designed to explore how advanced technology can support and amplify human decision-making in complex military scenarios.
AI vs. Human Planners: What the Experiment Measured
During the most recent DASH event, held in the fall of 2025, the Air Force brought together:
- AI tools from six different technology developers,
- Human planners from the U.S., Canada, and the U.K.,
- A set of multi-domain tactical challenges that included tasks such as planning airstrikes, rerouting aircraft after base damage, gathering intelligence on unusual signals, and defending a disabled Navy ship.
Each participant, both humans and machine,s was asked to produce Courses of Action (COAs): detailed tactical plans that address specific combat or operational problems.
The results were striking:
Speed
One of the AI systems delivered acceptable COAs up to 90% faster than human planners, dramatically shortening the time needed to generate options for complex military scenarios.
Accuracy & Quality
The top-performing AI produced solutions that were 97% viable and tactically valid, outperforming humans who achieved a much lower success rate with only 48% of human-generated COAs judged viable.
Volume of Options
AI systems were able to analyze vast amounts of data simultaneously and retain every detail from briefing documents, giving them an advantage in information-rich environments where humans may overlook or forget critical details.
Why These Results Matter for Future Warfare
The experiment’s success doesn’t mean computers will replace human commanders anytime soon. Rather, it points to a future where human judgment is augmented with machine speed and data-processing power, a concept the Air Force calls human-machine teaming.
In real combat situations, commanders must interpret rapidly changing information across multiple domains air, land, sea, space, and cyberspace. Analysts say that signal processing speed and pattern recognition abilities make AI especially well-suited for these environments, where the volume of data can overwhelm even the most experienced human planner.
Colonel John Ohlund, director of the Air Force’s Advanced Battle Management System Cross-Functional Team (ABMS CFT), stressed that AI tools “didn’t hallucinate,” a common AI risk where systems generate incorrect output in this particular experiment. He noted that this was achieved by training the AI on well-structured, military-validated datasets and aligning its outputs with trusted tactical information.
Challenges That Show AI Is Not Yet a Standalone Solution
Despite these impressive results, leaders caution that AI is not ready to replace human planners. None of the six AI systems tested are cleared for real-world operational use without further development.
Why AI Still Needs Human Collaboration
- Human judgment remains essential for understanding battlefield context and command intent.
- The experiment’s scenarios were run with unclassified datasets because real command-and-control networks and data used by the Air Force are classified and restricted, making direct integration a more complex task.
- Participants in the experiment were deliberately stretched outside their core areas of expertise, making some tasks unfamiliar even to seasoned planners a design choice meant to stress-test the human participants as much as the AI.
Even so, the AI’s performance digesting hundreds of pages of briefing documents and spreadsheets without losing critical details highlights how machine memory and rapid data processing could provide a strategic advantage in fast-paced operations.
What the Air Force Is Building Toward
The ultimate goal of the DASH program is not to have standalone AI battle planners, but to develop modular AI capabilities known as “microservices” that can plug into larger command-and-control systems. These systems will support human operators throughout the decision-making pipeline.
Within the Air Force’s broader Transformational Model for battle management, AI-generated COAs represent just one of multiple stages in developing an executable plan. Future experiments will explore how AI can assist with other critical steps, such as:
- Risk assessment and mitigation,
- Resource allocation across domains,
- Rehearsal and simulation of outcomes,
- Adaptive sequencing for campaign plans.
Lieutenant Ashley Nguyen, one of the experiment’s participants, summed it up well: “The AI didn’t replace us; it gave us a solid starting point to build from.”
AI as a Force Multiplier, Not a Replacement
Defense analysts say experiments like DASH are part of a broader push across the U.S. military to integrate AI into operational planning and execution. Artificial intelligence’s ability to process huge volumes of data rapidly and offer multiple viable tactical options could become a decisive advantage in future warfare particularly in multi-domain operations where speed and accuracy are paramount.
Yet experts also emphasize that while AI accelerates and expands planning capability, human oversight, critical judgment, and ethical control must remain central. AI must be paired with clear rules of engagement, robust validation procedures, and safeguards to prevent errors.
Looking Ahead: The Next Frontier in Human-Machine Teaming
The Air Force plans to continue its DASH experiment series through 2026 and beyond, each iteration refining how AI tools and humans collaborate under pressure. These trials aim to strengthen trust between operators and machine partners, fine-tune AI reliability, and accelerate integration into future command and control architectures.
As warfare evolves, the ability to pair human creativity and ethical judgment with machine speed and precision could be the key to maintaining decision advantage against peer rivals. This latest experiment underscores the potential for AI technologies to unlock new levels of operational effectiveness — transforming the art and science of battle planning in the process.
If AI can plan battles faster and more accurately than humans, at what point does human “oversight” become a liability rather than a safeguard?
Posts
0 Comments