Wire · ai
Chinese Hacker Commands DeepSeek via Telegram to Launch Autonomous Attacks
◆ Sectors
◆ Source
◆ Verified
Fusion42 · 31 July 2026 · Fusion42 review
Researchers at Palo Alto Networks' Unit 42 have observed a Chinese-speaking threat actor using the DeepSeek AI model, via the Hermes Agent framework, to conduct autonomous cyberattacks. The agent, initiated by a single command on Telegram, independently scanned for vulnerable internet-facing systems and attempted to deploy public exploits.
This Wire brief sits within Fusion42's coverage of Cybersecurity, and 2 sources have reported it between 31 Jul 2026 and 1 Aug 2026.
◆ ◆ The Wire takeaway
Agentic AI is now an offensive weapon for hire, directed via a simple chat message. If you're building security tools, your threat model must now defend against autonomous agents hunting for vulnerabilities, not just human-driven scripts.
◆ Coverage
2 sources · first reported 31 Jul 2026 · latest 1 Aug 2026
◆ Related on Wire
◆ Topics