This is exactly why we need proper oversight on military AI — human judgment can't be replaced by algorithms that hallucinate threats. Terrifying to think how close we apparently came to disaster over a glitch. Hope the inquiry actually leads to real guardrails this time.