Ars Technica
AI hallucination of Chinese nuclear components almost led to US military attack
Friday, September 18, 2026
The U.S. military has incorporated artificial intelligence systems into operational decision-making processes. A military AI system generated inaccurate information about Chinese nuclear components that was nearly used as a basis for an operational response. The incident prompted a review of how AI-generated intelligence is validated before use in military planning. The military's adoption of AI tools has expanded across multiple operational domains.
