Topic starter
02/08/2026 11:38 am
TL;DR: Agentic AI agents can call APIs, modify systems, and chain actions across tools, creating attack paths that traditional IAM, SIEM, and logging were not built to interpret, according to Panther. The real governance break is that human-paced review and static permission models assume the actor is reactive, not independently actioning tasks at runtime.
Editorial analysis by NHI Mgmt Group, based on content published by Panther: “Agentic AI Security Risks: What Changes When AI Can Take Action”.
Key questions
Q: What breaks when an AI agent can act inside a pipeline without human approval?
A: The assumption that workflow intent is fixed at definition time breaks first.
Q: Why do agentic AI systems create more security risk than standard chatbots?
A: Agentic systems can turn model output into action, which means a bad instruction can affect code flow, tool use, and downstream state.
Q: What are the signs that an agent’s effective permissions are too broad?
A: Look for agents that can move from retrieval to mutation, inherit user or service credentials without separation, or reach multiple environments with the same identity.
Practitioner guidance
- Scope agent identities to specific tasks Assign each agent a distinct identity with only the permissions needed for a defined task set, then review access by tool and by destination system instead of by application alone.
- Centralise agent telemetry in the SIEM Collect agent decisions, tool calls, token usage, and session identifiers in one pipeline so investigators can reconstruct the sequence behind an action instead of isolated local events.
- Require human approval for high-impact actions Put approval gates on outbound network requests, writes or deletes in production data, and any credential or permission change that the agent attempts during execution.
Bottom line: Agentic AI changes the security problem because an identity can now make runtime decisions, call tools, and move across systems in ways IAM was not designed to model.
Explore further
View Full Forum → | NHI Foundation Course → | Our Services → | Read the full analysis →
02/08/2026 12:51 pm
Agentic AI security exposes an assumption collapse in human-paced identity governance: access review processes were designed for privileges that persist long enough to be observed, certified, and revoked. That assumption fails when an agent can request, use, and discard access within a single task or execution path. The implication is not merely faster review cycles, but a rethink of where governance starts and what evidence exists to govern.
A few things that frame the scale:
- Only 13% of organisations feel extremely prepared for the reality of agentic AI despite the majority racing toward autonomous adoption, according to the 2026 Infrastructure Identity Survey.
A question worth separating out:
Q: How should organisations govern AI agents that can change production monitoring?
A: They should treat agent permissions as high-risk delegated access and require explicit scoping, auditability, and rollback. If an AI agent can change dashboards or alert rules, the organisation needs controls that limit scope drift and preserve human accountability for every production change.
👉 Read our full editorial: Agentic AI security risks are outpacing conventional IAM controls