Governance Crisis in AI Agent Security: CISOs Struggle for Control

A recent survey of over 300 CISOs reveals a governance crisis in enterprise security regarding AI agents. Many organizations lack visibility and control over these agents, with only 47% able to identify all AI agents in their environment. The fragmented governance approach, relying on shared credentials and broad permissions, exacerbates the issue. Additionally, only 31% of CISOs feel aligned with their C-suite and board on acceptable AI risk levels. The rise of AI agents in production workflows poses significant risks, particularly in regulated environments where data residency and compliance are critical. Organizations are urged to implement agentic identity and access management (IAM) to secure these non-human identities effectively. The report emphasizes the importance of treating AI agents as distinct identities with their own credentials and governance frameworks. Without proper management, organizations risk accumulating unmanaged automation and security vulnerabilities.

Key Points: • Only 47% of CISOs can identify all AI agents in their environment. • 31% of CISOs feel aligned with their C-suite on acceptable AI risk levels. • Organizations must treat AI agents as distinct identities with their own governance.