Palo Alto Networks’ Unit 42 threat intelligence team tracked a hacking campaign conducted by Chinese-speaking attackers between August and September 2026. The attackers breached 10 of at least 19 targeted organizations, including firms in the semiconductor and defense sectors and civil society groups.

The attackers used the open-source AI framework Cairn alongside Chinese-developed models, including Zhipu GLM and DeepSeek. After operators provided a target and a goal, the AI system planned and carried out intrusions. In some cases, it accessed valid credentials and created persistent administrator accounts.

At one company, the attackers exfiltrated more than one million records. Unit 42 said the campaign shows that AI-driven attacks can be deployed at scale by a small number of operators.