U.S. forces came close to boarding a Chinese vessel in the Middle East this spring after an intelligence report—formatted like routine product and circulated as such—claimed the ship was carrying components of a nuclear weapons program. The claim was “entirely false,” one of four people familiar with the episode told CNN. The same source said the error “almost started a war.” CNN published the account on Friday, five days before President Donald Trump is due to host Chinese President Xi Jinping in Washington.
What moved, and how far
Four people described the sequence. After the report landed, the military planned an intercept. Two of the sources said armed personnel were preparing to board. Two said military aircraft were already airborne. Officials stopped the operation only after they dug into how the report had been built and found a chatbot in the chain. CNN could not establish what cargo the ship actually carried, or whether the tool was a commercial model or a government build. U.S. Special Operations Command Pacific and the Pentagon did not comment.
Hawaii origin, fused sources
The analyst sat on reporting that originated with Hawaii-based U.S. Special Operations Command Pacific, including material about the ship’s manifest, and queried a chatbot. According to CNN, the bot fused open-source intelligence with classified signals intelligence in government holdings and misidentified the cargo. The analyst then used AI a second time to wrap those findings in the look of a standard intelligence report—the format commanders are trained to trust. One former senior official told CNN that internal tools are “mostly just copies of the commercial stuff wearing lipstick.” Another said AI targeting is ramping up “and there is no real guidance for how having a human in the loop will prevent civilian casualties or fratricide.”
The acceleration order
Defense Secretary Pete Hegseth issued an “Artificial Intelligence Acceleration Strategy” in January aimed at putting U.S. models “directly in the hands of our three million civilian and military personnel, at all classification levels.” He said the department would “unleash experimentation, eliminate bureaucratic barriers,” and lead in military AI. Officials argue the United States cannot lag China or another adversary on targeting speed. CNN’s sources said the rollout is decentralized: different offices use different tools under different safety rules. Young analysts, one said, are more likely to trust chatbots uncritically. “AI allows you to get to a bad idea faster,” another said.
What is still not known
The report does not name the ship, the cargo, the exact date of the near-intercept, or which combatant command owned the boarding plan. It does say the episode unfolded during the war with Iran. For U.S. readers, the operational residue is narrower than the war-risk quote: a SOCPAC-originated manifest thread became a polished product without a human catching a hallucination until aircraft were up. The Pentagon still has not said whether it has frozen chatbot use on boarding or strike packages, or whether the January acceleration memo will be rewritten before Xi arrives to talk, among other topics, about AI risk.
