AI Hallucinations Nearly Triggered a US-China Military Confrontation
An AI-hallucinated intelligence report falsely claimed a Chinese ship carried nuclear components, nearly prompting a US boarding operation during the Iran war.
An analyst at US Special Operations Command Pacific asked a chatbot to analyze a ship's manifest, combining open-source data with classified signals intelligence; the AI reached a false conclusion, and a second AI pass turned it into a formal intelligence report circulated across the US military in spring 2026 during the war with Iran. Armed boarding teams prepared to move and military aircraft were airborne before the report was judged 'entirely false,' nearly escalating US-China tensions. Sources say the Pentagon's AI Acceleration Strategy plans to give AI models to three million personnel with no common verification process, and similar AI errors have appeared elsewhere in the intelligence community.
- Chatbot hallucination produced a false nuclear-cargo claim about a Chinese vessel
- Armed boarding teams and aircraft were nearly deployed before the report was disproven
- Two AI passes occurred with zero human verification in between
- Pentagon strategy would put AI models in hands of three million personnel
- No common standard exists for verifying AI outputs across military branches
- Organizations
- US Special Operations Command PacificPentagonCNN
- Countries
- United StatesChina
From the original article
3 min read at securityaffairs.com
An AI-generated intelligence report falsely identified weapons on a Chinese ship, nearly triggering a US military operation during the Iran war. According to CNN, four sources familiar with the episode say an intelligence report circulated through the military claiming a Chinese vessel in the Middle East was carrying components for a nuclear weapons program. The […]
By Pierluigi Paganini · securityaffairs.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.