"United States Nearly Went to War with China Over an AI-Generated False Report," CNN Reports

[Financial News]  


CNN reported on the 18th (local time), citing sources, that the United States was nearly duped by a fake AI-generated report last spring and almost went to war with China.
According to the report, while the war in Iran was raging, an intelligence report quickly circulated within the U.S. military claiming that a Chinese vessel was transporting components related to a nuclear weapons program in the Middle East.
Four sources said the U.S. military immediately moved to intercept the vessel. Armed troops were ready to board it, and military aircraft were deployed overhead.
But just before the operation was carried out, officials closely reviewed the report and identified problems. They confirmed that an analyst with the Special Operations Command had written it with the help of AI and that the chatbot he used had misidentified the cargo aboard the vessel.
One source said the report was "completely false," but "war almost broke out." Any U.S. military operation against the Chinese vessel could have risked escalating into an armed clash between the two countries.
The report that nearly triggered an armed clash between the two superpowers, the United States and China, originated when an analyst asked a chatbot about intelligence concerning the cargo manifest of a Chinese vessel from the U.S. Pacific Special Operations Command. After combining publicly available information with classified signals intelligence from the U.S. government, the chatbot erroneously concluded that the cargo was linked to a nuclear weapons program.
The analyst then used AI to put the information into the format of a standard U.S. military intelligence report and circulated it internally.
A former senior U.S. official pointed out that the AI systems used by military and intelligence analysts are "mostly copies of commercial products with only the outer packaging changed" and can produce errors.
CNN noted that the case illustrates how the risk of military miscalculation is growing as the United States Department of Defense (DoD) and intelligence agencies rapidly expand their use of AI.
In January, Defense Secretary Pete Hegseth began expanding AI adoption across military operations, including intelligence analysis and target selection, under the "AI Acceleration Strategy." The aim is to gain an edge over adversaries through faster decision-making on the battlefield.
United States officials, however, voiced concern that government departments use different tools under different directives and safety standards, with no unified validation criteria, leading to major differences in reliability.
One source warned, "AI only gets you to bad ideas faster."

[email protected] Song Kyung-jae Reporter