TL;DR: Akeyless says federal agencies modernising infrastructure and adopting AI need identity controls that treat machine identities and AI agents as access-bearing actors, because authenticated access can still exceed task scope and expose sensitive systems, APIs, and data. Access governance must move from static credentials to runtime control and just-in-time privilege.
At a glance
What this is: Akeyless frames federal AI adoption as an identity security problem where machine identities and AI agents can be authenticated yet still over-privileged for the task at hand.
Why it matters: This matters because IAM teams now have to govern non-human and autonomous access patterns in environments where static credentials, broad permissions, and runtime tool use can all widen blast radius.
👉 Read Akeyless's discussion on AI agent identity security for federal agencies
Context
Federal agencies are expanding automation and AI in environments where machine identities and AI agents can touch sensitive systems, APIs, and data. The identity problem is not just authentication, but whether the granted permissions stay bounded to the task, the session, and the system being accessed.
For IAM, PAM, and NHI programmes, that changes the control point. The article positions identity-based, just-in-time access and runtime controls as the practical response to access paths that are no longer stable enough for credential-first governance models.
Key questions
Q: How can teams govern machine identities and AI agents in access reviews?
A: Teams should assign ownership, define review cadence, and include machine identities and AI agents in the same certification logic as human access, but with role-appropriate approvers. If a non-human identity can act on sensitive data, it needs a lifecycle owner and a removal path just like any other privileged account.
Q: Why do read-only AI agents still create serious security risk?
A: Read-only agents can still expose secrets, topology, environment variables, and other sensitive operational data. In practice, disclosure often creates the same downstream risk as modification because attackers can use the information for lateral movement, credential theft, or targeted follow-on attacks. Security teams should govern read-only access as a data-exposure path.
Q: What breaks when organisations keep using standing privileges for machine identities?
A: Standing privilege turns NHIs into persistent trust anchors. If the credential is embedded in code, reused across environments, or left valid after the workload changes, the access outlives the business purpose. That creates a durable attack path and makes it difficult to prove that the identity is still needed.
Q: When should organisations prioritise JIT access over persistent credentials?
A: Organisations should prioritise JIT access when the credential is only needed for a discrete task, session, or approval-bound workflow. If the access does not need to exist continuously, persistence adds risk without adding useful capability. JIT is most valuable where human admin access and machine workflows both create standing-credential exposure.
Background and context
Why static credentials fail for machine identities and AI agents
Static credentials create a durable trust relationship that often outlives the task, session, or approval that justified it. For machine identities, that means long-lived secrets, tokens, or keys can be reused well beyond the intended action. For AI agents, the problem is sharper: the agent can be authenticated yet still invoke actions, access data, or retrieve credentials beyond the task scope. In practice, the issue is not authentication alone, but whether authorisation is bounded tightly enough to the runtime context.
Practical implication: treat credential persistence as a governance risk, not just a secret-management issue.
How just-in-time access changes the identity model
Just-in-time access replaces standing privilege with time-bounded, task-scoped entitlement. Instead of granting a machine identity or agent a durable permission set, access is issued only when needed and withdrawn once the task ends. That reduces the window in which stolen or misused credentials can be leveraged. It also aligns better with Zero Trust Architecture, where access decisions are continuously re-evaluated rather than assumed stable after login or provisioning.
Practical implication: design access issuance around task completion, not account existence.
Runtime controls for AI agent actions
Runtime controls matter because an AI agent can remain properly authenticated while still becoming unsafe through tool calls, delegated actions, or chained execution. The control problem is to govern what the agent may do after authentication, not merely whether it can log in. That requires policy enforcement at action time, especially where agent behaviour can touch privileged workflows, secrets, or critical systems. In this model, the important boundary is not identity presence, but action authorisation at runtime.
Practical implication: enforce policy at action execution, not just at initial access grant.
NHI Mgmt Group analysis
Static credential governance breaks down when the actor is a machine or agent: the article is really about moving identity security from durable access to ephemeral, task-scoped control. That shift matters because the risk is not only credential theft, but permission overreach that persists after the original task context has changed. Federal teams should read this as a control-model change, not a feature discussion.
Just-in-time access is the decisive boundary for high-trust environments: agencies cannot keep relying on standing credentials when machine identities and agents are expected to touch sensitive systems. Just-in-time issuance narrows exposure and fits better with Zero Trust objectives, but only if entitlement removal is automatic and tied to task closure. The programme implication is that access duration becomes a first-class risk variable.
Agentic runtime governance is now part of identity governance: once an AI agent can initiate tool use or reach critical systems, identity control extends beyond authentication into action supervision. That brings AI agent identity under the same governance discipline as machine identity, while adding runtime policy enforcement as a distinct layer. The field is moving toward identity blast radius as a measurable control objective.
Vendor and federal context are converging on the same problem statement: agencies need security models that assume authenticated non-human access can still be excessive, misaligned, or too persistent. That is why the strongest control pattern here is not broader access, but narrower issuance and tighter runtime permissioning. Practitioners should expect NHI governance and agentic AI governance to merge operationally.
From our research library:
- NHIs outnumber human identities by 25x to 50x in modern enterprises, according to the Ultimate Guide to NHIs.
- 19% of organisations give AI systems dramatically more access than human employees, nearly one in five granting unrestricted privilege, according to the 2026 Infrastructure Identity Survey.
- Read next: Ultimate Guide to NHIs — Why NHI Security Matters Now
What this signals
Identity blast radius: agencies should measure how far a machine identity or AI agent can move if its access is misused, not just whether it can log in. That means mapping permissions to task boundaries, secret exposure paths, and privileged action points before automation scales further.
Runtime controls are becoming the control plane for AI adoption in federal environments. Authentication remains necessary, but it is no longer sufficient when an agent can use tools or touch critical systems after login.
Just-in-time issuance and strict secret isolation are the two levers most likely to reduce non-human access risk without slowing adoption. The programme question is whether identity governance can keep pace with automation without reverting to standing privilege.
For practitioners
- Define task-scoped access for non-human actors Map each machine identity and AI agent to a single business task, system boundary, and expiration condition. Eliminate broad standing entitlements where the access case can be expressed as a time-bounded work item.
- Move privileged access to just-in-time issuance Issue privileged credentials only at the moment of need and revoke them automatically when the approved action completes. Do not leave reusable credentials embedded in agent workflows or service operations.
- Block credential exposure inside agent workflows Prevent AI agents from seeing, storing, or relaying secrets that are not required for their task. Keep credentials outside the agent execution path wherever possible and broker access through controlled runtime policy.
- Enforce runtime controls on agent actions Apply policy checks at the point where an agent tries to invoke a tool, reach a system, or trigger a privileged action. Treat the execution decision as the security boundary, not the authentication event.
Key takeaways
- Machine identities and AI agents should be governed as access-bearing actors, not as simple authentication events.
- Static credentials and standing privilege expand exposure when non-human actors operate across sensitive systems and APIs.
- Just-in-time issuance and runtime controls shift governance from durable access to task-bounded access decisions.
Standards & Framework Alignment
This section maps relevant standards and security frameworks to the operational risks and controls described in this guidance.
OWASP Non-Human Identity Top 10 addresses the attack and risk surface, while NIST CSF 2.0, NIST Zero Trust (SP 800-207) and NIST SP 800-53 Rev 5 set the governance and control requirements practitioners need to meet.
| Framework | Control / Reference | Relevance |
|---|---|---|
| OWASP Non-Human Identity Top 10 | NHI-04 — Insecure Authentication | The article centres on authenticated non-human access that can still overreach task scope. |
| NHI-05 — Overprivileged NHI | The core issue is access that exceeds what a machine identity or agent needs for the job. | |
| NHI-07 — Long-Lived Secrets | The article argues against static credentials and toward short-lived access for non-human actors. | |
| Recommendation — Harden non-human authentication so valid access cannot be reused beyond the intended task. Reduce NHI entitlements to the minimum scope needed for each workload or agent action. Replace durable secrets with short-lived credentials and automatic expiry for non-human actors. | ||
| NIST CSF 2.0 | PR.AA-05 — Access Permissions, Entitlements and Authorizations | The subject is the control of permissions for machine identities and AI agents. |
| Recommendation — Map non-human permissions to PR.AA-05 and verify entitlement scope against task requirements. | ||
| NIST Zero Trust (SP 800-207) | Principle 3 — Never trust, always verify | The article aligns with continuous verification for non-human access in Zero Trust environments. |
| Recommendation — Apply continuous verification so non-human access is checked at use time, not only at login. | ||
| NIST SP 800-53 Rev 5 | IA-5 — Authenticator Management | The article depends on secret and credential lifecycle control for machine identities and agents. |
| Recommendation — Use IA-5 to govern credential issuance, expiry, and revocation for non-human identities. | ||
Key terms
- Machine Identity: The digital identity of a machine, device, or workload, such as a server, container, or VM, used to authenticate it within a network. Sometimes used interchangeably with NHI, though NHI is the broader category.
- Just-in-Time Access Request: Just-in-Time Access Request is a pattern that grants access only when it is needed and only for the duration required. It reduces standing privilege by making access temporary, policy driven, and task scoped. This approach is especially useful for contractors, sensitive systems, and short-lived operational work.
- Runtime control: Controls that enforce policy while an AI system is operating, rather than after the fact. For healthcare chatbots, runtime control includes data masking, output filtering, access scoping, and immutable logging so the organisation can defend the interaction itself.
- Standing Privilege: Standing privilege is access that remains active even when no immediate task requires it. For NHI programmes, it is a common failure mode because long-lived credentials and persistent roles create unnecessary exposure. Reducing standing privilege usually means tighter expiry, on-demand access, and clearer review of who or what still needs access.
What to expect at the briefing
Akeyless's full discussion covers the operational detail this post intentionally leaves for the source:
- Live discussion of how federal teams can apply identity-based just-in-time access in regulated environments
- Speaker perspectives on keeping credentials out of AI agents while still supporting secure automation
- Practical framing for runtime controls that govern autonomous agent actions before they reach critical systems
- Context on how FedRAMP Class D environments shape identity security decisions for government teams
Deepen your knowledge
NHI governance, agentic AI identity, and machine identity security are core topics in our NHI Foundation Level course, the industry's only accredited NHI security programme. If you are building or maturing an IAM programme, it is worth exploring.
Published by the NHIMG editorial team on October 13, 2026.
NHI Mgmt Group, the independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org