Bottom line: The Tenable RSO team's tracking of agentic AI threat activity has revealed a concerning trend, indicating a significant increase in near-autonomous offensive AI capabilities.
What's happening: The Taiwan autonomous AI cyber attack in August 2026, which targeted Taiwan's Ministry of Defense, was part of a cluster of seven incidents tracked by Tenable's RSO team since late July 2026. These incidents involved attacks on multiple countries, including the United States, China, and Indonesia, using tactics, techniques, and procedures (TTPs) commonly associated with near-autonomous AI. The attackers exploited vulnerabilities in software from two major vendors, including a known CVE-2022-45350 vulnerability in a widely-used operating system, resulting in a CVSS score of 9.8.
What to do: Security leaders should review their incident response plans and ensure that their threat detection systems are equipped with the latest AI-powered threat intelligence to identify and respond to agentic AI threats. Additionally, they should prioritize patching all known vulnerabilities, including the CVE-2022-45350 vulnerability, and implementing robust security controls to prevent future attacks.