AI CRUSHES Human Military Planners – Unbelievable!

(LibertySociety.com) – AI just crushed human military planners in a high-stakes battle simulation, generating viable war plans 90% faster with zero errors—what does this mean for America’s fighting edge?

Story Snapshot

Air Force’s DASH-3 experiment in September 2025 pitted AI against U.S., Canadian, and UK planners in complex battle scenarios.
AI produced courses of action up to 97% viable, far surpassing humans’ 48% rate, in under a minute versus 19 minutes per plan.
No AI hallucinations occurred due to structured data; humans struggled under stress and information overload.
Col. John Ohlund leads effort, stressing AI as human enhancer for decision superiority against peer threats.
Results signal shift to human-AI teaming in Joint All-Domain Command and Control (JADC2).

DASH-3 Experiment Details

U.S. Air Force conducted DASH-3 at Shadow Operations Center-Nellis in Las Vegas during September 2025, including September 24. Six unnamed companies supplied AI tools. Human teams from U.S., Canada, and UK generated courses of action (COAs) for scenarios like airstrikes, aircraft rerouting, intelligence gathering, and protecting a disabled Navy vessel. AI outperformed across speed, volume, accuracy, and data retention under stress.

Humans averaged 19 minutes per COA with 48% viability and frequent errors. One AI tool delivered 97% viable, tactically valid plans. AI generated more options up to 90% faster, considering risk, fuel, and geospatial factors. Experiment stressed participants with novel, multi-domain challenges outside routine expertise.

Col. John Ohlund, ABMS Cross-Functional Team director, noted AI’s perfect recall: “The computer does not forget.” Lt. Ashley Nguyen called AI a “solid starting point,” user-friendly for time savings. No tools replace humans; they augment under pressure.

Evolution from Prior DASH Experiments

DASH series stems from Advanced Battle Management System (ABMS) to modernize command-and-control. DASH-2 in July 2025 saw AI generate 30 times more options in under 10 seconds, producing over 6,000 solutions hourly with human-level accuracy after tweaks[2]. DASH-3 advanced to coalition teaming and complex COAs in minutes.

Air Force Research Lab’s 711th Human Performance Wing, Integrated Capabilities Command, and 805th Combat Training Squadron hosted. Focus: microservices integration into systems for peer conflicts like those with China or Russia.

November 2025 capstone tested joint AI battle management. National Academies highlight ABMS/JADC2 as prime AI proving grounds. Iterative improvements eliminated early subtle errors.

Implications for Military Superiority

Short-term, AI accelerates training and stressed operations, cutting planning time dramatically. Long-term, hybrid human-AI command transforms JADC2, enabling faster multi-domain decisions vital for deterrence. Vendors eye contracts; allies gain interoperability.

Conservative values prioritize warfighter strength—facts show AI reduces human error under duress, aligning with common-sense preparation against threats. Ohlund’s caution on data prep rings true: success demands human oversight, not blind reliance. Social concerns over job shifts exist, but national security trumps.

AI’s zero hallucinations via structured inputs set precedent for high-stakes use. Ohlund and Col. Jonathan Zall emphasize coalition human-machine teaming. No operational deployment yet; evolution targets risk analysis and multi-kill chains.