Dark models: AI is making Chinese cyberattacks harder to detect and punish
In mid-September, Donald Trump announced the creation of a new agency called AI Force — his response to statements from industry leaders about the need to slow the pace of the technology’s development. However, opponents of implementing a voulntary slowdown are fearful of the effects such a step could have on competition with China, whose technology is becoming increasingly popular among hacker…
In July meetings across Taiwan, the National Security Bureau reported 2.4 million intrusion attempts daily against government networks in 2024, and 2.63 million in 2025. Despite the seemingly pressing nature of Chinese cyberattacks and a ballooning cyber campaign, the government's response was slow. China's espionage through its networks involves industrial-scale espionage, and artificial intelligence is entering the contest between Washington and Beijing, altering who can detect the attacker.
A tool from Israeli firm Dream demonstrated near-autonomous AI attacks on Taiwanese government systems, compromising accounts and exfiltrating personnel records. DeepSeek and other open-source models were used to generate exploit code and harvest IP addresses, highlighting AI's role in cyberattacks. The U.S. Justice Department's seizure of domains behind QScan and QTRouter, built by a contractor group QTFY linked to China's Ministry of State Security, further illustrates the threat.
While AI can simplify and accelerate known cyberattack methods, it cannot independently discover new vulnerabilities.
Written by urgent.news from The Insider's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
