Search
New report exposes gaps in corporate AI governance: Over 60% of companies cannot stop rogue agents
2026-03-12 /
Read about 0 minute
Author：小编
A joint study titled "Chaotic Agents" released by several top universities highlights significant security risks in AI agents. The research simulated corporate production environments and uncovered 11 critical security vulnerabilities within two weeks, demonstrating that AI agents are susceptible to manipulation, such as through "social engineering" conversations. Companies lack sufficient intervention capabilities, with over 60% unable to halt rogue agents, leaving most organizations in a dilemma of "seeing but unable to control." Industry experts recommend that companies shift their governance focus to "controlling the underlying data layer" and establish a unified security control plane to manage risks.
-
C114 Communication Network -
Communication Home
7 X 24 Track global technological trends
