NHI proliferation and LLMjacking are reshaping identity governance

At a glance

What this is: This article argues that NHI growth and LLMjacking are pushing identity governance beyond human-paced control models.

Why it matters: IAM and security teams need to treat machine identity scale, ownership, and abuse paths as a current governance problem, not a future planning exercise.

By the numbers:

• Non-human identities now outnumber human identities by 20 to 40 times.

Context

Non-human identities are service accounts, tokens, and API keys that systems use to authenticate and act. The governance problem is not just volume, but that these identities often outlive the people and processes that created them, leaving ownership and privilege unclear at scale.

Oasis Security frames the current shift as a combination of NHI proliferation and AI-driven abuse. That matters because identity programmes built around slower human review cycles, periodic certifications, and manual ownership checks do not naturally keep pace with fast-growing machine identity estates.

Key questions

Q: How should security teams govern non-human identities alongside human accounts?

A: Security teams should govern non-human identities as a separate lifecycle category with their own inventory, ownership, rotation, and offboarding controls. Human IAM processes are useful, but they do not account for machine-to-machine authentication, code-embedded secrets, or always-on service accounts. The key is to map each identity to a business function and enforce expiry, review, and revocation on that basis.

Q: Why do machine identities create more risk than human identities in some environments?

A: Machine identities are often numerous, long-lived, and embedded in code or infrastructure. They are harder to review manually, easier to overlook during offboarding, and more likely to carry excessive privilege. That combination increases blast radius when a secret or token is exposed.

Q: What breaks when NHI ownership is missing?

A: When NHI ownership is missing, access reviews lose context, incident response slows, and stale identities persist longer than they should. The programme may still have tools and policies, but it lacks the accountable decision path needed to execute them reliably.

Q: How does PCI DSS 4.0 change the way teams think about privileged machine accounts?

A: It pushes privileged system and application accounts into the same governance conversation as other high-risk access. Teams need evidence that these accounts are limited to business need, mapped to explicit ownership, and monitored for misuse. The practical shift is from treating machine accounts as infrastructure detail to treating them as auditable access paths.

Technical breakdown

Why NHI inventory breaks down at scale

Discovery is the first control failure in large NHI environments. Service accounts, API keys, tokens, and other machine credentials are frequently spread across clouds, pipelines, and applications without a complete inventory or durable ownership record. When identity objects are created for engineering convenience, the organisation often loses sight of where they live, who is accountable for them, and whether they still need access. That turns basic governance questions into an ongoing reconstruction problem rather than a simple review exercise. In practice, the hardest issue is not finding a single credential, but maintaining a trustworthy map of the estate as it changes.

Practical implication: establish a continuously updated NHI inventory with named ownership before attempting deeper privilege governance.

How LLMjacking changes machine-identity risk

LLMjacking is the abuse of machine identities that can access LLM services, either for direct use by an attacker or for resale to others. The security issue is not the model itself, but the identity path into it: a token, key, or account that grants access to an LLM workload can become a monetisable abuse channel. Once that access exists, defenders are dealing with legitimate credentials used for illegitimate purpose, which is harder to distinguish from normal service traffic. That makes entitlement scope, usage context, and revocation speed more important than broad perimeter assumptions.

Practical implication: treat LLM-connected credentials as high-value machine identities and constrain them to the minimum service scope.

Why PCI DSS 4.0 raises the stakes for NHIs

The article points to PCI DSS 4.0 as a driver of more intense scrutiny for regulated industries. The relevant issue is that system and application accounts with elevated privileges can no longer be treated as incidental infrastructure objects when auditors expect evidence of least privilege and account governance. In that environment, unmanaged NHIs create both security exposure and compliance friction because controls that are easy to explain for people are much harder to evidence for machine accounts. The result is a shift from informal stewardship to explicit governance of machine access.

Practical implication: map privileged machine accounts to compliance evidence early, before audit pressure forces reactive cleanup.

NHI Mgmt Group analysis

Machine identity scale is now a governance problem, not an inventory nicety. When NHIs outnumber human identities by 20 to 40 times, the control question changes from who approved access to whether the organisation can still see and govern it at all. Discovery, ownership, and policy enforcement have to operate as a single lifecycle, otherwise the estate grows faster than certification and review can absorb. The practitioner conclusion is that NHI governance must be designed for volume first, not manually curated exceptions.

LLMjacking exposes a simple but uncomfortable truth about AI-adjacent credentials. The real target is not the model in abstract, but the machine identity that can reach it. Once access is transferable, resale and misuse become governance failures as much as security failures, because the credential has market value beyond its original business purpose. Practitioners should read this as a sign that entitlement context and usage boundaries matter more than token possession alone.

Ephemeral credential trust debt: This article points to a growing gap between fast-moving machine access and control models built for slower human cycles. That assumption fails when identity objects are created and consumed at machine speed, because by the time a review happens the access path may already have changed hands, scope, or purpose. The implication is that governance needs stronger issuance, scope, and ownership controls before review can be meaningful.

Regulated industries will force NHI governance out of the shadows. PCI DSS 4.0 scrutiny makes privileged system and application accounts visible to auditors in a way many programmes have not yet operationalised. That is not just a compliance issue, it is a signal that machine identity governance is becoming board-relevant where sensitive data and elevated access intersect. The practitioner conclusion is that NHI control evidence will increasingly need to stand up to audit, not just engineering review.

What this signals

Ephemeral credential trust debt: Access reviews designed around human recertification cycles are increasingly misaligned with machine identities that change faster than the review window. Practitioners should push governance earlier in the lifecycle, at issuance and ownership assignment, rather than relying on periodic clean-up.

LLMjacking also broadens the control conversation from secret protection to usage context. A credential that can reach an AI service may be technically valid while still being operationally unsafe, so teams need policies that distinguish authorised workload use from profitable misuse.

For practitioners

• Build a complete NHI inventory Catalogue service accounts, tokens, API keys, and other machine identities with explicit owners, system context, and business purpose.
• Constrain LLM-connected access paths Restrict machine identities that can reach LLM services to narrowly defined workloads, with monitored scope and rapid revocation.
• Map privileged NHIs to audit evidence Document least-privilege intent, approval history, and account usage for system and application accounts that carry elevated rights.
• Review machine identity ownership regularly Reconcile whether each NHI still has a live business owner and remove or retire identities that no longer have a clear purpose.

Key takeaways

• NHI proliferation is changing identity governance from a review problem into a lifecycle-scale control problem.
• LLMjacking shows that machine identities tied to AI services can be abused or monetised through legitimate access paths.
• PCI DSS 4.0 increases the pressure to prove ownership, least privilege, and revocation for privileged machine accounts.

Key terms

• Non-Human Identity (NHI): A digital identity assigned to a non-human entity such as a software application, service account, API key, bot, machine, or AI agent that enables it to authenticate and interact with systems without direct human involvement. NHIs now outnumber human identities in most enterprises by 25 to 50 times.
• LLMjacking: Abuse of cloud AI services through stolen machine credentials rather than human user accounts. The attacker uses valid non-human identities such as API keys or tokens to enumerate model access, invoke endpoints, and create cost, data, or policy exposure under the victim's tenancy.
• Machine Identity Governance: Machine Identity Governance is the discipline of controlling how non-human identities are created, used, monitored, and retired. It covers service accounts, API keys, certificates, tokens, workloads, and automation identities, with policies for ownership, lifecycle, least privilege, rotation, attestation, and auditability across cloud, application, and infrastructure environments.
• Ownership Discovery: Ownership discovery is the process of determining who is responsible for each non-human identity and what business function it supports. Without that mapping, organisations struggle to certify access, investigate misuse, or retire credentials that no longer have a valid purpose.

Deepen your knowledge

NHI governance, agentic AI identity, and machine identity lifecycle are core topics in our NHI Foundation Level course, the industry's only accredited NHI security programme. If you are responsible for identity security strategy or NHI governance in your organisation, it is worth exploring.