Topic starter
03/02/2026 4:23 pm
TL;DR: Thousands of likely AI identities in customer environments can be identified by a new tagging feature, including more than 2,800 at one large enterprise, by combining name heuristics, trust policy signals, and usage analysis, according to Token Security. The real issue is not visibility alone but the identity assumptions that fail when AI activity hides behind ordinary service accounts.
Editorial analysis by NHI Mgmt Group, based on content published by Token Security: “Tagging Identities with AI Access: Token Security Finds Shadow AI in Your Enterprise”.
By the numbers:
- On launch, the feature identified more than 2,800 likely AI identities at one large enterprise.
Key questions
Q: What breaks when AI use hides inside ordinary service accounts and app registrations?
A: The identity model breaks because teams can no longer tell whether a credential represents a human workflow, standard automation, or AI-driven access.
Q: Why do shadow AI tools create identity governance risk?
A: Shadow AI is risky because users often reach those tools through identities, browser sessions, or tokens that were never assessed for data handling or access scope.
Q: How do security teams decide which identity data needs stronger controls?
A: Start by asking whether the data helps an attacker choose who to target next.
Practitioner guidance
- Map AI-accessible identities from multiple signals Correlate naming patterns, trust policies, attachments, and usage telemetry so AI-driven access is not inferred from a single weak signal.
- Review service accounts and connected apps for AI reach Check whether ordinary service accounts, connected apps, and API keys are being assumed or invoked by AI platforms, then document the reason each identity was tagged.
- Tie tags to access review and monitoring workflows Use the AI_ACCESS reason code to route identities into heightened monitoring, scoped entitlement review, and lifecycle decisions.
Bottom line: Shadow AI becomes an IAM blind spot when AI activity rides through identities that look ordinary to human reviewers.
Explore further
View Full Forum → | NHI Foundation Course → | Our Services → | Read the full analysis →
This topic was modified 6 days ago by NHI Mgmt Group
06/10/2026 6:34 pm
Shadow AI identity tagging exposes an identity interpretation failure, not just a discovery gap: the problem is that existing IAM catalogues treat service accounts, roles, and connected apps as ordinary machine identities even when AI is actively using them. That means the governance issue begins before remediation, at the point where teams decide what an identity object represents. The practical conclusion is that identity state and identity behaviour are no longer the same thing.
A few things that frame the scale:
- 70% of organisations grant AI systems more access than they would give a human employee performing the exact same job, according to the 2026 Infrastructure Identity Survey.
- Systems with least-privileged AI access had a 17% incident rate vs 76% for over-privileged systems. Organisations failing to scope AI access properly are 4.5x more likely to experience a security incident, according to the 2026 Infrastructure Identity Survey.
A question worth separating out:
Q: Should organisations treat AI_ACCESS tagging as a replacement for access reviews?
A: No. Tagging is a discovery and triage control, while access review is a governance control. The tag tells you where AI-related access may exist, but teams still need entitlement validation, monitoring, and lifecycle decisions to reduce the risk.
👉 Read our full editorial: Shadow AI identity tagging exposes enterprise access blind spots