The ability to attribute each AI-driven data access event to a specific actor, dataset, and policy outcome. It is the audit standard privacy and security teams need when a non-human actor can query regulated records, because token validity alone does not prove what was actually read.
Expanded Definition
Who-accessed-what governance is the control discipline that links each AI-driven retrieval to the actor, the target dataset, and the policy decision that allowed or denied access. It goes beyond token issuance or session authentication and asks what was actually queried, when, under which purpose, and with which entitlements in force. In NHI environments, that distinction matters because an autonomous agent can hold a valid credential yet still access records outside the intended scope.
This concept sits at the intersection of auditability, data minimization, and non-human identity oversight. The OWASP Non-Human Identity Top 10 treats over-privilege, secret sprawl, and weak lifecycle controls as core risk patterns, while the NIST Cybersecurity Framework 2.0 frames traceability as part of governance and detection readiness. Guidance varies across vendors on how much query-level context is enough, but no single standard governs this yet.
The most common misapplication is treating authentication logs as proof of governed access, which occurs when teams record only token use and not the underlying dataset, policy outcome, or downstream query scope.
Examples and Use Cases
Implementing who-accessed-what governance rigorously often introduces logging overhead and privacy review complexity, requiring organisations to weigh forensic clarity against operational cost and data minimization obligations.
- An AI agent queries a regulated customer file, and the log records the agent identity, the exact dataset, the policy rule that permitted the lookup, and whether row-level filters truncated the result set.
- A finance workflow invokes a service account through Ultimate Guide to NHIs — Lifecycle Processes for Managing NHIs and later proves which invoice records were read during reconciliation.
- A vendor-connected automation accesses shared CRM data, and the security team uses Ultimate Guide to NHIs — Regulatory and Audit Perspectives to support evidence collection for audit and legal review.
- A support chatbot triggers sensitive account lookup, and investigators correlate the access path with the incident pattern documented in the Meta AI Instagram Account Takeover case study.
- An internal data assistant is constrained by policy and identity bindings, then validated against the logging expectations in the NIST SP 800-53 Rev 5 Security and Privacy Controls.
Why It Matters in NHI Security
When who-accessed-what governance is missing, organisations can prove that an identity existed without proving what it saw, which weakens incident response, privacy assurance, and regulatory defensibility. That gap is especially dangerous in AI and NHI environments where a single credential may be reused across tools, data domains, and automated tasks. It also makes least-privilege enforcement difficult because access reviews cannot distinguish legitimate narrow use from silent overreach.
NHIMG research shows the scale of the problem: in The 2024 ESG Report: Managing Non-Human Identities, 72% of organisations reported experiencing or suspecting an NHI breach, and inadequate monitoring is a recurring weakness in the State of Non-Human Identity Security. That is why audit teams increasingly ask for evidence that aligns with the scope of access, not just the success of authentication. Organisational exposure becomes visible only after an investigation, subpoena, or breach review, at which point who-accessed-what governance is operationally unavoidable to establish.
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 address the attack and risk surface, while NIST CSF 2.0, NIST SP 800-63, NIST Zero Trust (SP 800-207) and NIST AI RMF set the governance and control requirements practitioners need to meet.
| Framework | Control / Reference | Relevance |
|---|---|---|
| OWASP Non-Human Identity Top 10 | NHI-02 | Addresses excessive NHI access and weak visibility into what non-human actors can reach. |
| NIST CSF 2.0 | DE.CM-8 | Supports monitoring that captures user and device activity across assets and data. |
| NIST SP 800-63 | AAL2 | Identity assurance alone does not prove what a credentialed actor actually accessed. |
| NIST Zero Trust (SP 800-207) | PA | Policy enforcement should verify each request before data is released to an entity. |
| NIST AI RMF | AI risk management requires traceability of model and agent actions impacting data. |
Log and review dataset-level access so every AI action can be tied to a specific identity and policy outcome.
