Skip to content
CORVASEC — Knowledge Defends
NIST / CPRT

NIST AI 100-2e2025

A NIST publication describing adversarial machine learning concepts, attack types, and potential mitigations. AI teams can use its terminology and examples when assessing threats to models, data, and AI-enabled systems.

NIST Dataset Global public
OVERVIEW

About this reference

NIST AI 100-2e2025 is a dataset reference associated with adversarial machine learning taxonomy and mitigations. It sits within nist / cprt and is most useful when its guidance or terminology is connected to a clearly defined system, risk question, or organizational requirement. The publisher document, rather than this editorial summary, establishes the authoritative scope and wording.

Primary purpose

Use this reference when work involves adversarial machine learning taxonomy and mitigations. Its relevance depends on the source’s stated scope and the environment being assessed.

Subject areaNIST / CPRT
Intended readersSecurity, risk, governance, and technical teams
PRACTICAL GUIDANCE

Applying this reference

Suggested steps for applying this material. Adapt these to your environment and confirm them against the primary source.

  1. Confirm source dataset, revision, and import date before using control identifiers.
  2. Use identifiers to crosswalk frameworks and highlight gaps without treating a mapping as proof of compliance.
  3. Apply the relevant parts of NIST AI 100-2e2025 to a scoped example, recording any assumptions and open questions before adoption.

Review checkpoints

  • Locate the authoritative NIST AI 100-2e2025 publication and establish its version and scope.
  • Determine what adversarial machine learning taxonomy and mitigations means for the relevant assets, systems, or processes.
  • Record interpretation questions, proposed actions, and where the primary source supports them.
ASSESSMENT & EVIDENCE

Validation and documentation

These examples can support an implementation review; they are not a publisher-defined audit checklist.

  • Retain original identifiers, mappings, evidence locations, and decisions about implementation.
  • Maintain a record of the document edition or update date, applicability decisions, and associated source links.

Considerations and limitations

  • Crosswalks are aids to analysis; they do not replace the underlying publication or evaluation.
  • Verify whether the source has been amended, superseded, withdrawn, or replaced before treating any requirement as current.
SOURCE & REVISION

Publisher and source details

Publisher
NIST
Category
NIST / CPRT
Document type
Dataset
Applicability / jurisdiction
Global
Access
public
Catalog ID
nist-ai-100-2e2025
Catalog snapshot
2026-09-24

Before relying on this reference

  1. Confirm the current edition, errata, and publication status.
  2. Verify licensing and whether the primary text is publicly available.
  3. Check applicability to the organization, industry, jurisdiction, and system.
  4. Record the source URL and the date it was reviewed.
View authoritative publication ↗