AI Hallucination

Policy and Re

2026-09-20 03:43:45

CNN says U.S. military nearly boarded a Chinese ship after AI-driven intelligence error

CNN reported that during the U.S.-Iran war this spring, the U.S. military came close to launching an armed boarding operation against a Chinese vessel in the Middle East after intelligence supported by AI wrongly identified the ship as carrying components for a nuclear weapons program. According to four people cited by CNN, aircraft were already in the air and armed personnel were preparing to board before officials revisited the report at the last minute and found the conclusion was wrong. The report says the intelligence came from an analyst at U.S. Special Operations Command Pacific in Hawaii, who used a chatbot to examine cargo manifest information. The chatbot reportedly combined open-source material with classified signals intelligence and concluded that the ship was transporting nuclear-related components. CNN said it could not determine what the vessel was actually carrying, only that the chatbot had misidentified the cargo. The episode has added to scrutiny of how AI tools are being introduced into military targeting and intelligence work. CNN also cited sources who said similar hallucination-style problems are not isolated cases inside the intelligence system, especially as analysts face pressure to produce and distribute assessments more quickly.

290

GPTZero

2026-07-24 08:05:17

EY Canada 44-Page Report 72% AI-Generated: 16 of 27 Citations Fabricated, GPTZero Exposes 'Citation Poisoning'

GPTZero found that 16 out of 27 citations in an EY Canada cybersecurity report were AI hallucinations, with ~72% of the text AI-written. The report contained contradictory figures and fake sources; EY has since retracted it.

180