Urgent.News

What's breaking now, across thousands of outlets.

AI

China’s J-36 fighter jet designers report danger of military AI hallucinations

Large language models are expert in intelligence work, able to quickly mine key data and analyse weapons performance – tasks that once took analysts huge amounts of time and effort. But engineers at China’s AVIC Chengdu Aircraft Research and Design Institute, developer of the People’s Liberation Army’s backbone J-20 and next generation J-36 stealth fighters, have urged caution. In a paper…

China’s J-36 fighter jet designers report danger of military AI hallucinations

Chinese engineers working on the J-36 fighter jet have expressed concerns over the potential dangers of artificial intelligence hallucinations in military intelligence work. Zhang Xianzhe, an engineer from China's AVIC Chengdu Aircraft Research and Design Institute, which developed the J-20 and J-36 stealth fighters, warned in a June 20 paper that AI models can fabricate aircraft specifications and inaccurate radar data, which could lead to strategic miscalculations.

Every fighter jet must undergo a meticulous intelligence gathering process, considering factors such as enemy radar coverage, missile no-escape zones, electronic warfare capabilities, and ground-based air defense networks. Zhang cautioned that large language models, not trained on authoritative military sources, often produce plausible but false outputs, including incorrect materials, aerodynamic impossibilities, and fictional military elements.

These hallucinations could have severe consequences in defence intelligence, potentially causing false assessments of combat capability. A real-world example cited was the 2026 US strike on Iran, which mistakenly targeted a school due to outdated data and an AI system flagging it as a high-priority target. Zhang suggested several mitigation strategies, including using trustworthy data sources, cross-checking AI recommendations, and promoting safer applications of large language models in defence technology.

Written by urgent.news from Reuters Business via SCMP's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at scmp.com →

More in AI

Stop treating AI like it's human

AI doesn't need human motives to produce autonomous harm. If people use such breaches for personal gain, the motive is human, even if AI systems facilitate the action.

A year after their last raise, Germany’s Flip raises €22 million to scale its frontline AI platform

Flip, a Stuttgart-based employee experience platform for frontline workers, has raised €22 million (£19 million) in order to fund AI infrastructure and knowledge development for frontline roles.

  • Germany's Flip raises €22 million for AI platform expansion
  • Led by Notion Capital and HV Capital, investment round includes L-Bank and Baden-Württemberg bank
  • Flip aims to bridge AI gap in frontline roles with secure digital identity and custom apps

More from Wednesday 19 August →