September 3, 2025 — Leads & Copy — Delinea released research highlighting gaps in AI security. Only 44% of organizations believe their security architecture is fully equipped to secure AI, despite 93% expressing confidence in their machine identity security, according to a survey of over 1,700 IT decision-makers worldwide.
The report, titled “AI in Identity Security Demands a New Playbook,” reveals that while 61% of organizations have full visibility into all machine identities for monitoring, only 48% have identity governance for AI entities. With 66% of organizations utilizing agentic AI, limited visibility poses risks such as unchecked autonomous decisions and potential compromise.
Delinea CEO Art Gilliland emphasized the need for adaptive, risk-aware systems. Key findings show that 94% of global companies use or pilot AI in IT operations, with over half using generative or agentic AI. However, 56% face shadow AI issues monthly, and only 57% have acceptable use policies for AI tools.
Organizations identified AI-generated phishing, credential theft, unchecked agentic AI systems, unsanctioned AI usage, and poor visibility into AI access workflows as top concerns. The report underscores the importance of robust identity governance and adaptive security controls.
Justin Ordman, Corporate Communications Director, can be reached at PR@delinea.com.
Source: Delinea
