Connect
A Q3 2025 IT Trends Report reveals a stunning statistic—99.6% of organizations are already using or actively planning to adopt artificial intelligence.
AI is being embedded everywhere, from security threat detection (67%) to AI-powered helpdesks (68%). While this innovation is exploding, a silent security gap is widening.
We are so focused on the human users of AI that we have overlooked the new users it creates. The single biggest unaddressed risk in this new AI-powered world is the explosive, ungoverned growth of non-human identities.
The Ghost in the Machine: What Is a Non-Human Identity?
Non-human identities, also known as agentic AI, are not just user accounts; they are the AI tools themselves. These identities are the scripts, bots, API keys, and autonomous “agents” that AI systems use to communicate with other systems, access data, and execute tasks. The risk here is significant, as these agents often require elevated privileges to function.
These agents connect to sensitive databases, critical applications, and privileged accounts, all without a human in the loop. This isn’t a theoretical problem. The report shows that 42% of IT leaders are already concerned about a “lack of identity governance for non-human actors,” and this fear is growing.
The Great Disconnect: 42% Worry, Only 23% Act
Here lies an alarming disconnect. While nearly half of IT leaders express concern, the same report shows only 23% of IT teams are actively securing these agentic AI identities. This gap exists for a few key reasons.
Traditional identity and access management (IAM) was built for human beings, not machines. These identities are often “invisible,” buried in code or API settings rather than appearing in a user directory. Furthermore, 51% of leaders cite “AI tools integrating with sensitive systems without proper review” as a top risk, showing that things are moving too fast for current security practices.
This creates a perfect blind spot for security teams. An unsecured, unmonitored AI agent becomes a wide-open, privileged backdoor into your organization’s most critical systems.
The Path Forward: From IT Blocker to Enabler
IT leaders are caught in a difficult position. The business wants to innovate, as shown by the 99.6% adoption rate. However, a staggering 94% of IT leaders are rightfully concerned about the new vulnerabilities AI creates.
Saying “no” to AI is not the answer, as the business will find ways to work around IT, creating even more “shadow AI.” The only way to lead is to adopt AI intentionally. This requires a new governance model that manages all identities—both human and machine—from a single, unified platform.
AI is now an essential part of nearly every organization’s strategy, and the security stakes have never been higher. As non-human identities proliferate, your attack surface grows—often in ways that are hidden from traditional controls.
The time to act is now; every delay increases the vulnerability of your most critical assets.
Taking decisive steps to secure non-human identities does more than fill a security gap—it protects business continuity and builds trust with stakeholders.
To build a framework that secures your entire environment and allows you to enable AI safely, you need a new playbook. Download the Q3 ’25 IT Trends Special Report | AI Edition to learn how you can get started today.
94% of IT leaders worry about vulnerabilities introduced by AI.
Get the latest trends and advice on the state of AI in this IT Trends special report.
