Glossary · simply explained

Least privilege

Least privilege means: every identity gets only the rights it really needs — and only for as long as necessary.

The least privilege principle demands that every identity — human, service or machine — receives exactly the rights necessary for its task: no more, no longer, no broader. It is one of the oldest security principles and at the same time the most frequently violated.

The reason for its effectiveness is simple: the damage of a compromised account is exactly as large as its rights. Minimal rights mean a minimal blast radius.

Why rights sprawl in practice

Privilege creep happens gradually: project changes bring new access, old access is never revoked; template copying (like colleague X) multiplies over-provisioning; under time pressure rights are granted generously because revocation never presses anyone. After years nobody has the overview — until an incident forces it.

The remedies are processes instead of appeals: granting rights via roles (RBAC) instead of individual approvals, just-in-time rights for administrative work, automatic revocation on change and exit, regular recertification — and visibility into which rights are actually used.

Least privilege beyond identities

  • Network: segmentation and ZTNA grant access to individual applications instead of whole networks.
  • Workloads: microsegmentation restricts which systems may talk to each other.
  • Cloud: roles with a narrow scope instead of accounts with wildcard rights.
  • Applications: API keys and tokens with minimal scope and expiry.

Frequently asked questions about Least privilege

What does the least privilege principle state?

Every identity receives exactly the rights its task requires — no more, no broader, not permanently. That keeps the damage of a compromised account or system limited to the minimum the task would have required anyway.

How do you introduce least privilege realistically?

Step by step and data-driven: first create visibility (who has what, who uses what), then remove the biggest risks — admin rights, orphaned accounts, broadband roles. Role models, just-in-time access and recertification keep the state stable afterwards.

What is privilege creep?

The gradual growth of permissions across job changes and projects because new rights are added but old ones never revoked. Long-serving employees accumulate access profiles no role model ever intended — a favourite target for attackers.

Does the principle apply to machines and services too?

Especially there: service accounts, API keys and workload identities are numerous, long-lived and rarely monitored. Minimal scopes, short lifetimes and clear owners are the most effective protection — also between systems, via microsegmentation.

How does least privilege relate to Zero Trust?

Least privilege is one of the pillars of Zero Trust: every access is verified and granted only to the minimally necessary extent. Zero Trust architectures such as ZTNA implement the principle at the network level — access to the single application instead of the whole network.

More terms

LLM (large language model)

Large language models process and generate language — the base technology behind ChatGPT, copilots and enterprise AI.

Lateral Movement

Lateral movement refers to techniques with which attackers work their way from system to system through a network after the initial breach. The goal is to expand privileges, find sensitive data, and ultimately bring critical systems under control.

Malware

Malware is the umbrella term for software that deliberately causes harm: from data theft to the encryption of entire systems. The main types differ in their distribution path and goal, with direct consequences for the appropriate defense strategy.

From term to implementation: KAEMI supports you from the first assessment to the ongoing managed service.