Researchers observe first ‘near-autonomous’ AI attack on government target in Taiwan

Threat actors deployed an AI-powered offensive framework capable of autonomously discovering vulnerabilities, pivoting across targets, and self-correcting during a live attack against the Taiwanese government and its IT supply chain. The root issue is that traditional, human-paced defenses are increasingly mismatched against AI-accelerated attack cycles that can adapt faster than conventional incident response allows. The lateral expansion into energy companies and IT vendors demonstrates how a single intrusion can cascade into critical infrastructure compromise when supply chain relationships are insufficiently secured. This incident signals a paradigm shift in threat velocity — organizations must assume attackers can now operate at machine speed, making continuous monitoring and automated response no longer optional.