Air Force Cadet Builds ROMAD-AI Prototype in 3 Months Using 3 Chatbots
Updated
Updated · MIT News · Jul 7
Air Force Cadet Builds ROMAD-AI Prototype in 3 Months Using 3 Chatbots
1 articles · Updated · MIT News · Jul 7
Summary
Three months of prompt-based coding let U.S. Air Force cadet Joshua Lynch build ROMAD-AI, a prototype app developed with Claude, ChatGPT and Gemini despite his having no prior coding experience.
The project, run through the Air Force-MIT AI Accelerator's Phantom Program, tested whether nontechnical service members could turn military problem knowledge into usable software without the traditional development pipeline.
ROMAD-AI was initially aimed at battlefield support, including target recognition and communications, but Lynch scaled it back to document processing—analyzing tactical maps and generating mission-planning documents—as AI limits and development time emerged.
The prototype showed chatbots can speed nonexpert prototyping, but not safely replace full development for sensitive or critical uses; Lynch found one version was sending documents to Gemini for analysis instead of processing them locally.
MIT Lincoln Laboratory mentor Laura Niss said the effort highlighted AI's value as a bridge between operational experts and technical teams, while code review and security still remain major bottlenecks.
With novices building military apps via AI, what is the new, indispensable role for human experts?
As AI writes half of all new code, how can the military prevent a catastrophic security failure?
Empowering Novice Coders: The ROMAD-AI Prototype and the Future of Secure, Ethical AI Integration in U.S. Military Operations
Overview
The ROMAD-AI prototype, led by U.S. Air Force cadet Joshua Lynch under the Department of the Air Force–MIT AI Accelerator Phantom Program, was created to democratize AI development for military applications. By empowering nontechnical service members to build functional AI programs, the project aimed to bypass traditional, lengthy development pipelines and unlock the potential of the military’s broader workforce. Using a novel 'vibe-coding' approach with generative AI chatbots, ROMAD-AI enabled non-coders to contribute directly to technological advancements, making AI development more accessible and impactful for both battlefield and support operations.