TeamT5 says AI tools have more than doubled attacks by China-linked hackers

Taiwan cybersecurity company TeamT5 said attacks by Chinese state-backed hacking groups have risen by more than 100% after those actors began using AI tools to write malicious code and scan networks. The company specifically named DeepSeek among the models being used, and said the same groups have also used OpenAI’s ChatGPT and Anthropic’s Claude Code. The report, cited by Techub News and attributed to The Decoder, points to a shift in how threat actors are operating: AI models are being applied directly to offensive tasks such as code generation and network reconnaissance. TeamT5’s warning centers on the pace of that change rather than on a single campaign. The item also referenced a UK study that found open-source models are rapidly catching up in cyberattack capabilities. Together, the two points suggest that offensive AI tooling is becoming more accessible across both commercial and open model ecosystems, even as the source note did not disclose details on the specific timeframe or campaigns involved.