AI-driven cyberattacks target government networks and global servers

New reports highlight a significant shift in cyber warfare as threat actors increasingly utilize artificial intelligence to automate and adapt attacks. In the APAC region, specifically Taiwan, a near-autonomous AI attack framework has been deployed against government networks. This framework moves beyond simple automation to agentic operations, using LLM-driven feedback loops to autonomously rewrite code and modify malware signatures in real-time to bypass security detections.

Simultaneously, a Chinese-speaking hacker group identified as UAT-10147 has been observed using AI to target approximately 170,000 vulnerable Windows and Linux servers globally. This group targets diverse sectors, including government agencies, universities, and technology firms, across countries such as Brazil, China, Canada, and Vietnam. The group employs sophisticated tools like the SPECTRE backdoor and custom rootkits, with researchers noting that some components may have been constructed with AI assistance to enhance methodical execution and evasion capabilities.