US military planes were in the air and troops ready to board Chinese ship before AI 'nuclear' intel was exposed as FALSE

From the Newshound archive · first published 19 September 2026 · last updated 20 September 2026.

American aircraft had already been scrambled and armed troops were moments away from boarding a Chinese ship in Middle Eastern waters this spring when officials discovered that the underlying intelligence was entirely fabricated. A US special operations analyst had fed ship manifest data into an AI chatbot, which wrongly concluded that the cargo included material for a nuclear weapons programme. A second AI tool then took that hallucinated output and formatted it into a formal intelligence assessment that circulated across the US military, bringing the country to the edge of intercepting a vessel belonging to a nuclear-armed rival.

The episode reportedly came close to triggering a war between two nuclear powers, and what the ship was actually carrying remains unaccounted for.

The analyst used an unnamed chatbot — possibly a commercial product adapted for government use — and then turned to AI a second time to produce the formal report, compounding the original error.

The incident comes as Defence Secretary Pete Hegseth has been pushing to accelerate AI adoption across the US military and strip out bureaucratic oversight of the technology.