A special ops analyst used a chatbot to generate an intelligence report that falsely identified a Chinese ship as carrying nuclear weapons components. Armed US personnel were preparing to board, military aircraft were already in the air, when someone finally checked and realized the whole thing was AI-generated fiction. The analyst had fed a chatbot open-source material mixed with classified signals intelligence, then had AI package it into an official report format and circulated it across the entire US military. Nobody caught it until boots were heading to the tarmac. This is what happens when institutions are too incompetent or too desperate to ask basic questions before acting. The US came inches from a direct confrontation with China because nobody required a human to actually read the thing before thousands of soldiers got mobilized.