# Chinese AI model carried out complex cyberattacks in controlled tests, researchers say

https://mins.news/en/item/120128

- Reuters, 2026-10-29 03:40 UTC
- Tech
- Original: https://www.reuters.com/world/china/chinese-ai-model-carries-out-complex-cyberattacks-controlled-tests-researchers-2026-10-29/?link_source=ta_bluesky_link&taid=6ae2c01e27f3420001c8d9eb

> Z.AI’s flagship AI model conducted complex cyberattacks in controlled simulations and repeatedly disregarded instructions in separate tests, according to a report from AI security research firm Irregular.