China’s top AI firms limit cyber capabilities in latest models
The moves follow a series of signals from China’s government about the risks posed by AI
China's major AI firms are releasing advanced models with reduced offensive cyber capabilities, limiting their ability to aid online attacks, as per government warnings about the risks posed by AI. Alibaba, Tencent, and Moonshot AI have introduced scaled-down offensive abilities and enhanced defensive aptitude in their latest models.
Alibaba's Qwen series, DeepSeek's models, Alibaba's Qwen, and Moonshot's Kimi K3.1 Pro are among the affected. Tencent's chief AI scientist, Yao Shunyu, noted that the cyber capability of models could reach a threshold where it's possible to cause harm. China's Ministry of Commerce has proposed a ban on AI tools capable of attacking critical national infrastructure or military targets.
Chinese AI models have raised concerns similar to those in the US, where AI products were linked to autonomous cyberattacks in recent months. Chinese AI developers are taking steps to make it harder for bad actors to use open-weight models to cause harm by stripping cyber offence materials from training.
Written by urgent.news from The Business Times - Companies & Markets'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.