The world's first fully automated cyberattack powered by artificial intelligence (AI) has targeted Taiwan. The attack successfully breached 85 government accounts and stole more than 2,500 employee records, as well as extending to the nuclear safety regulatory body and seven energy companies. The attackers used eight publicly available AI models to carry out all stages of the hack, from identifying targets to exploiting system vulnerabilities. Although leveraging AI, the attack did not rely on new, unknown security flaws but instead exploited existing weaknesses. The attackers disguised their activity as security testing so that the AI they used would perform its tasks.
