The Department of Defense has launched a new generative artificial intelligence tool that will help the agency identify and engage military targets, according to a DOD source and an explosive leak reported on by journalist Jessica Burbank and Drop Site News.
Defense Secretary Pete Hegseth announced the new AI tool – dubbed “Gemini” – in a Tuesday memo obtained by Burbank, informing all DOD personnel that they are expected to incorporate the technology into their “workflows immediately.” Hegseth wrote that Gemini would help the agency “out-think, out-decide, and out-pace any adversary.
According to a source within the DOD who spoke with Burbank on the condition of anonymity, the rollout of Gemini also included the distribution of a “Do and Don’t list.”
“[One] example of don’t was, don’t use GenAI for decisions involving attribution, targeting, or threat evaluation without human validation; I read that as someone can read what the AI said and be like ‘yep it’s good to go shoot that missile,’” the source told Burbank, who published the report on her Substack, "Ka's Counterprogramming." “They are legit going full force into AI.”
Hegseth would officially announce the technology Tuesday morning in a social media post on X, though the details – including its use for helping to identify military targets – were not revealed until Burbank and Drop Site News’ reporting later Tuesday morning. Instead, Hegseth highlighted the technology’s ability to “conduct deep research, format documents and even analyze video or imagery at unprecedented speed.”
“Victory belongs to those who embrace real innovation,” wrote Hegseth in the internal memo, issued to all DOD personnel.
“Rather than being reliant on the dusty, antiquated systems of a bygone era, we are thinking ahead here in the Department of War. [Gemini] is part of this monumental transformation. It removes wasted time and focuses more of our energy into decisive results for the warfighter.”