Taiwan says AI agents used in July cyberattacks

Taiwan's digital affairs ministry said overseas hackers used artificial intelligence agents to help carry out cyberattacks targeting Taiwanese government agencies in July.

The ministry said the affected agencies had completed response measures, but did not give details on the scale of the attacks, the damage caused or the agencies targeted. It did not identify a country behind the latest incident.

The ministry said hackers used a hybrid model that combined AI agents, including the open-source AI agent platform OpenClaw. It said AI agents can quickly link multiple attack techniques and use backup or testing systems as launch points, giving attacks speed, low cost and scale.

The statement followed a Financial Times article that identified Taiwan as the target of an autonomous AI attack by suspected Chinese hackers. The Financial Times, citing researchers at Israeli AI company Dream, said hackers used open-source AI agents to build an autonomous tool that mapped 21 government systems, explored vulnerabilities and changed tactics when blocked.

Kenny Huang, chairman of Taiwan Network Information Center, said AI cyberattacks were not new, but that this was the first time it had been revealed that multiple AI agents were used to carry out a cyberattack.

Taiwan has previously blamed Chinese hackers for most of the millions of cyberattacks affecting the island each day. China claims Taiwan as part of its territory, while Taipei accuses Beijing of using cyberattacks and other “greyzone” tactics to pressure the island. China has previously denied hacking allegations and said it opposes all forms of cyberattacks.