CNN reports the US military nearly intercepted a Chinese ship based on an “entirely false” AI-generated report.

According to investigative reports, the United States military narrowly avoided a major international confrontation in the Middle East after nearly intercepting a Chinese vessel based on erroneous information.

Armed personnel and military aircraft were deployed and prepared to board the ship following an official intelligence briefing suggesting the vessel was transporting components destined for a nuclear weapons development program. However, moments before the critical operation commenced, officials examined the source of the intelligence compiled by a special operations analyst. They discovered the report was built with the assistance of an artificial intelligence chatbot, which had completely fabricated the details regarding the cargo.

This alarming incident is part of a broader pattern of failures linked to the integration of automated technologies in defense operations. A separate investigation by the Pentagon into a deadly missile attack on an Iranian school during the winter also pointed to a dangerous overreliance on artificial intelligence systems as a primary contributing factor to the failure. These instances highlight the severe risks of incorporating experimental automated models into military workflows. The reliance on unverified software outputs in volatile situations raises critical ethical and operational concerns, especially when machine learning hallucinations have the potential to spark armed conflict or cause civilian casualties.