Subscribe to the Non-Human & AI Identity Journal
Authentication, Authorisation & Trust

CIBA

← Back to Glossary
By NHI Mgmt Group Updated June 12, 2026 Domain: Authentication, Authorisation & Trust

CIBA is a backchannel approval pattern for situations where the user is not in a browser at the point of decision. In identity programmes, it gives teams a way to capture out-of-band approval for sensitive actions without forcing the agent or application to fake a normal interactive login flow.

Expanded Definition

CIBA, or Client-Initiated Backchannel Authentication, is an OpenID Connect pattern used when the user is not present in the browser or front end at the moment a sensitive decision is made. The client initiates authentication, but the approval step happens through a separate channel, often a mobile device, authenticator app, or other out-of-band workflow. The OpenID Foundation describes the model in the CIBA specification, which matters because identity teams should treat it as a protocol pattern rather than a generic approval message. See the OpenID Connect CIBA Core and the NIST Cybersecurity Framework 2.0 for the governance context around verifying access events.

In NHI and agentic AI programmes, CIBA is useful when an autonomous workflow needs human approval without pretending to be a browser-based user session. It is especially relevant for high-risk actions such as spending approvals, privileged configuration changes, and cross-system data release. Definitions vary across vendors when CIBA is folded into broader “step-up auth” or “push approval” language, so teams should keep the protocol boundary clear: CIBA is about backchannel initiation and asynchronous user confirmation, not a replacement for robust policy, device trust, or privilege control. The most common misapplication is using CIBA as a shortcut to justify weak session design, which occurs when teams treat the approval prompt as sufficient proof of intent while ignoring context, replay risk, and entitlement scope.

Examples and Use Cases

Implementing CIBA rigorously often introduces latency and user coordination overhead, requiring organisations to weigh stronger approval assurance against slower execution paths.

  • An AI agent prepares a payment release, then triggers a backchannel approval to a finance approver before the transfer is executed.
  • A service account requests elevation for a maintenance task, and the approver confirms on a separate trusted device rather than through the application session.
  • A healthcare workflow asks for patient-data export approval after the request is initiated by a backend process, reducing the need for interactive login fakery.
  • Security teams use CIBA-style approval for sensitive administrative actions where the browser is unavailable, but the approver must still validate the request context.

NHIMG’s Ultimate Guide to NHIs notes that 97% of NHIs carry excessive privileges, which makes approval patterns like CIBA useful only when paired with entitlement minimisation and auditability. In practice, CIBA is strongest when combined with policy checks, just-in-time access, and a trusted out-of-band channel, rather than used as a stand-alone control. The OpenID Connect CIBA Core remains the primary protocol reference for implementation details.

Why It Matters in NHI Security

CIBA matters because NHI and agentic systems often need to obtain human approval without a conventional user session, and that is exactly where attackers exploit confusion between authentication, authorization, and workflow routing. If a platform cannot prove who approved what, and under which conditions, the organisation loses defensible control over high-impact actions initiated by agents, service accounts, or delegated automations. This is especially important where approval is used to compensate for excessive privilege or weak segregation of duties. The NHI Management Group reports that 80% of identity breaches involved compromised non-human identities such as service accounts and API keys, underscoring why approval flows must be designed as part of identity governance, not as a cosmetic front-end step. For broader operational mapping, the NIST Cybersecurity Framework 2.0 helps align approval handling with access control, monitoring, and response expectations, while the Ultimate Guide to NHIs provides the NHI risk context.

Organisations typically encounter the need for CIBA only after an automated action has already been challenged, blocked, or disputed, at which point the approval trail becomes operationally unavoidable to address.

Standards & Framework Alignment

This section maps relevant standards and security frameworks to the operational risks and controls described in this guidance.

OWASP Agentic AI Top 10 address the attack and risk surface, while NIST CSF 2.0 and NIST AI RMF set the governance and control requirements practitioners need to meet.

FrameworkControl / ReferenceRelevance
OWASP Agentic AI Top 10Agent approval flows are a core concern for controlling autonomous action and human confirmation.
NIST CSF 2.0PR.AC-1CIBA supports controlled access decisions when users are not directly in-session.
NIST AI RMFAI risk governance covers human oversight for automated actions needing approval.

Require explicit approval boundaries before an agent can execute sensitive backchannel actions.

NHIMG Editorial Note
Reviewed and updated by the NHIMG editorial team on June 12, 2026.
NHI Mgmt Group — the #1 independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org