News / Korea / yonhap_finance

Chinese Hackers Increasingly Using DeepSeek AI for Cyberattacks Due to Low Costs and Weak Safeguards

Technology Services · Internet Software/Services · yonhap_finance · 2026-08-25

Research firm TeamT5 reports that Chinese hacking groups are leveraging the DeepSeek AI model to scale attacks, citing its low cost and minimal security.

What Happened

Rise in AI-Powered Cyberattacks: Taiwanese research firm TeamT5 reported that Chinese hacking groups are increasingly utilizing open-source AI models, specifically DeepSeek, to conduct cyberattacks. Since adopting AI, these groups have reportedly doubled the frequency of their operations.

Preference for DeepSeek: Hackers favor DeepSeek because it offers high performance with relatively weak security guardrails compared to other models. While more powerful alternatives like Moonshot's Kimi K3 exist, their high operational costs make them less attractive for malicious use.

Tactical Deployment: The report highlights that hackers use DeepSeek for various stages of an attack, including reconnaissance, vulnerability scanning, and the generation of malicious code. Specific groups have already been linked to using the AI to target corporate email systems and gather network data.

Security Vulnerabilities: Analysts note that the lack of strict safety protocols in Chinese AI models provides a significant advantage to state-linked hackers. This contrasts with Western AI models, which have more robust guardrails that are harder to bypass for illicit activities.

Read original