Hegseth Puts His Face on AI Tool That Labels His Strikes Unlawful

A new Pentagon-built artificial intelligence tool, GenAI.mil, has concluded that a hypothetical U.S. “double strike” on a drug-smuggling vessel — including a second attack on survivors in the water — would be “unambiguously illegal.” The scenario put to the system mirrors recent U.S. military operations in the Caribbean overseen by Defense Secretary Pete Hegseth, where boat strikes have killed at least 87 people and triggered bipartisan concern on Capitol Hill. In one of the most disputed incidents, on September 2, U.S. forces launched a follow-up strike on two survivors who were floating on debris.

GenAI.mil, introduced as an AI assistant for service members, has become a signature project for Hegseth, who is promoting it through a campaign echoing the “I Want You” poster — but with his own likeness. When asked whether ordering a second missile strike on shipwrecked survivors would violate U.S. military rules, the AI stated plainly that such an action would breach the laws of armed conflict and must be refused.

Its conclusion arrives just as Congress intensifies oversight of the real operations the scenario resembles. The

Senate Armed Services Committee recently opened a bipartisan investigation into the boat strikes, focusing in part on whether the September 2 mission violated longstanding protections for hors de combat — individuals who are wounded, incapacitated, or otherwise no longer participating in hostilities. Hegseth has rejected claims that he ordered the second strike but has acknowledged that the operation occurred.

These questions are unfolding against the backdrop of diminished legal oversight within the Department of Defense. In February 2025, President Trump and Hegseth dismissed several JAG attorneys, calling them “roadblocks,” a move legal analysts say removed critical internal safeguards against unlawful military action. At the same time, the administration is pursuing an Executive Order aimed at preventing states from regulating AI, even as it promotes the Pentagon’s own system to service members.

GenAI.mil’s determination aligns with what military lawyers and former commanders have repeatedly argued: attacking shipwrecked survivors is illegal under both U.S. policy and international law. The fact that the Pentagon’s own AI assistant reached the same conclusion now deepens questions about how the military’s actual conduct compares to the standards it claims to uphold.