'Almost Started a War': False AI-Linked Intel Triggered US Military Moves Towards Chinese Ship

American military aircraft were already airborne and armed personnel were preparing to board a Chinese vessel in the Middle East when officials discovered that an intelligence report produced with help from a chatbot had incorrectly identified the ship's cargo.

The episode, reported by CNN on Friday 18 September 2026 by Katie Bo Lillis and Zachary Cohen on the basis of four sources familiar with it, unfolded this spring during the war with Iran.

An intelligence report circulated across the US military claiming the ship was carrying components of a nuclear weapons programme, and preparations for the interception were halted only when officials looked more closely at how the assessment had been put together. One source described the report as entirely false and said it almost started a war.

How Close the Operation Came

The report set off an immediate response. According to CNN's account of the incident, the claim that a Chinese ship in the Middle East was transporting nuclear weapons components prompted the military to make plans to intercept the vessel, and two of the four sources said armed service members were preparing to board it.

Aircraft were already in the air. Two sources told the network that military planes were airborne when officials dug deeper into the report and established how it had been assembled, at which point the operation was stopped.

What made it dangerous was the target. CNN's sources described the incident as a potential close call that could have led to a direct confrontation with China, since an American operation against a Chinese vessel carried a serious risk of escalation.

It comes days before Chinese President Xi Jinping is scheduled to meet Donald Trump in Washington on 24 September, with artificial intelligence expected to feature among the issues under discussion.

What the Chatbot Got Wrong

The assessment was put together by a Special Operations Command analyst. CNN reported that it was produced with the help of artificial intelligence, with the analyst feeding the chatbot a mixture of classified signals intelligence and open-source material, from which it drew its conclusion about the ship's cargo.

The conclusion was incorrect. The chatbot inaccurately identified the material the vessel was carrying, and CNN was unable to establish what the ship was actually transporting.

Two basic questions remain unanswered. It is not clear whether the analyst used a commercially available chatbot or a government system such as GenAI.mil, and CNN could not establish which tool was involved.

CNN's reporting also does not identify any disciplinary action arising from the incident, and one source told the network that similar AI-related problems had occurred before, suggesting the case may not be isolated.

The department declined to address the specifics. In a statement quoted by Mediaite, the Department of War said it does not comment on specific intelligence reports, alleged operational planning or ongoing military assessments, while adding that it remains committed to the rigorous verification of all intelligence inputs and the responsible integration of emerging technologies.

The Pentagon's Push to Put AI Everywhere

The incident lands in the middle of a deliberate expansion. Officials across the military and the intelligence community are working to weave artificial intelligence into nearly every part of their work, from processing intelligence and supporting targeting operations through to budgeting, logistics and supply chains.

The spending has followed. In July 2025 the Defense Department awarded Google Public Sector a contract with a ceiling of about £147 million ($200 million) to accelerate artificial intelligence and cloud capabilities, and on 12 January 2026 Defense Secretary Pete Hegseth announced an Artificial Intelligence Strategy for the Department of War covering warfighting, intelligence and enterprise operations.

A dedicated platform arrived in between. Hegseth launched GenAI.mil, a US government AI platform, on 9 December 2025, with Google's Gemini for Government as its first capability. He said it would allow personnel to conduct deep research, format documents and analyse video or imagery at what he called unprecedented speed. Whether that platform had any role in the false report has not been established.

The timing sharpened the risk. The report circulated while the United States was fighting Iran, a conflict in which China's role has been repeatedly scrutinised.

CNN reported in April that sources familiar with US intelligence believed Beijing was preparing to ship man-portable air-defence systems to Iran through third countries, an allegation the Chinese foreign ministry denied, and the United States has separately sanctioned networks involved in procuring missile-propellant ingredients from China for Iranian weapons programmes.

The episode is one of the clearest publicly reported examples of an AI-assisted intelligence error reaching the point of triggering preparations for a military operation.

The ship sailed on, and CNN was unable to determine what was actually in its hold.