U.S. military AI for battlefield planning

In early July 2026 a U.S. Air Force cadet began creating a battlefield‑planning application that combined large‑language models such as ChatGPT, Claude and Gemini. The effort was an isolated experiment and did not yet involve formal military testing.

Later that month the U.S. Army’s Combat Capabilities Development Command introduced an experimental generative‑AI platform called COA‑GPT at the Project Convergence Capstone 6 exercise at Fort Irwin, California. The tool is designed to compress mission‑planning timelines from days or hours to seconds, allowing commanders to generate and compare courses of action through an interactive map interface. Developers stress that COA‑GPT is intended to augment, not replace, soldiers, and the system is being refined through continued soldier testing and industry partnership with Hyssos Tech. The deployments highlight growing use of AI in modern warfare and raise ethical questions about accountability for machine‑generated targeting decisions, as no international treaty currently governs autonomous weapons.