Trend Micro Flags AI Agent Governance Risks, Urges Automated Oversight

As agentic AI gains the ability to independently break down tasks, invoke tools and execute workflows, attackers could also use it to generate exploits at scale and automate reconnaissance and intrusions. Yuan Ze University professor Jen-Fu Wang said companies must govern not only their own systems but also open-source models, third-party components and data sources. The EU AI Act has also made traceability, accountability and risk controls key compliance issues.

Speaking at CYBERSEC Taiwan, Wang said AI agents and open-source models were reshaping the corporate threat landscape. He recommended establishing systematic, automated processes for vulnerability inventories, patching, verification and continuous monitoring, while bringing the model supply chain under governance. Available information about the event does not specify the date of the speech or cite any investment, loss or fine. The latest focus is on the need for companies to strengthen both technical defenses and AI accountability frameworks.