77% of IT managers say their AI agents are out of control - 5 ways to rein in yours (www.zdnet.com)

A recent survey by Rubrik ZeroLabs reveals alarming trends in AI agent management, finding that only 23% of IT managers feel they have complete control over these systems. More than 80% of respondents indicated that managing these agents requires more time than they save, and concerns are growing about security, with a significant number anticipating that the proliferation of AI agents will outpace the effectiveness of existing security measures within six months. This "agent sprawl" mirrors early cloud adoption issues, leading to inconsistent governance and heightened security risks. The report emphasizes the critical need for organizations to treat agent management as a primary discipline rather than an afterthought. Key recommendations include improving agent observability through comprehensive audit trails and employing telemetry to track agent actions effectively. As organizations prioritize speed in deploying AI systems, they must balance this against stringent governance to mitigate risks associated with autonomous operations. Without robust oversight mechanisms, companies face potential chaos as independent teams create countless agents with overlapping permissions and inadequate security configurations.