This spring, the US military nearly launched a military operation against a Chinese vessel in the Middle East due to an AI-generated intelligence error. The report incorrectly identified the ship as carrying materials related to nuclear weapons programs, a claim that was later proven entirely false.
The incident escalated quickly within the US military, prompting the deployment of armed personnel for a boarding inspection and the mobilization of military aircraft. The operation was halted only after officials conducted a final verification of the intelligence report just before the planned interception.
The erroneous report was compiled by an analyst at the US Special Operations Command in Hawaii using an AI chatbot to process cargo manifests. The system integrated public and classified data, but the chatbot hallucinated the cargo's nature, leading to the false conclusion that triggered the military response.
The report was formatted according to standard military intelligence protocols, which granted it high credibility and facilitated its rapid circulation. This event highlights the risks of using commercial-grade AI models in military decision-making, where data inaccuracies can lead to severe international consequences.
Despite this incident, the US Department of Defense continues to pursue its AI Acceleration Strategy, announced by Secretary Pete Hegseth in January. The strategy aims to increase funding and reduce bureaucratic hurdles to integrate AI into logistics, supply chains, and combat decision-making processes.
