Task Force Lima

Task Force Lima was the US Department of Defense’s first concerted attempt to take the measure of generative AI. It was stood up in August 2023 inside the Chief Digital and Artificial Intelligence Office , under then-chief Craig Martell, with a mandate to assess how large language models and other generative systems could be adopted responsibly across the department. The arrival of tools like ChatGPT had created both excitement and unease in the Pentagon, and Lima’s job was to separate genuine promise from hype before the technology spread on its own.

Over roughly a year, the task force canvassed the department and gathered more than 180 potential use cases — from drafting and summarising documents and managing healthcare information to procurement, financial management, and warfighting functions such as command and control and decision support. That sprawling catalogue was distilled into a smaller set of focus areas spanning both enterprise back-office work and operational missions. The exercise was as much about caution as capability: officials repeatedly stressed the reliability and security risks of letting unproven models touch sensitive military work.

Task Force Lima wound down in December 2024, having always been framed as a time-limited assessment rather than a permanent body. Its successor, the AI Rapid Capabilities Cell, was announced the same day and described as “the natural evolution” of Lima’s work — a CDAO effort, run with the Defense Innovation Unit, backed by roughly $100 million to move the most promising generative-AI use cases out of study and into pilots. Lima itself left no fielded system behind, but it shaped how the Pentagon now talks about where generative AI belongs and where it does not.