alt Bern
|
alt Lisbon
|
alt New York
+351 93 832 8533
alt Bern
|
alt Lisbon
|
alt New York
+351 93 832 8533

NIST Generative AI Profile Assessment

Confabulation, harmful content, data leakage, content provenance and third-party generative AI in your value chain.

NIST Generative AI 30 questions About 15 minutes Free, no account
Step 1 of 2 Organisation context
Organisation context

This shapes the recommendations you get at the end. Nothing here changes your score.

Please enter your organisation name.

Please enter a valid email address.

Please choose an option.

Please choose an option.

Please choose an option.

Please choose an option.

Choose another framework

Self-assessment only. It is not legal advice, an audit opinion or certification, and completing it confers no conformity status.

Readiness questions

Answer all 30 questions as honestly as you can. There are no right or wrong answers; an honest picture produces a useful action plan.

01

GenAI Governance and Policy

01 Is there an acceptable-use policy governing generative AI use by employees and in products?

Critical requirement, weight 3

02 Do you maintain an inventory of generative AI systems, models, and use cases (including shadow/unofficial use)?

Critical requirement, weight 3

03 Are approval processes in place before new GenAI use cases go live?

Critical requirement, weight 3

04 Are staff trained on GenAI limitations, risks, and safe usage practices?

Important requirement, weight 2

02

Pre-Deployment Testing and Red-Teaming

05 Are GenAI applications tested against defined criteria before deployment (quality, safety, robustness)?

Critical requirement, weight 3

06 Is structured red-teaming or adversarial testing performed (prompt injection, jailbreaks, misuse scenarios)?

Critical requirement, weight 3

07 Are evaluations repeated after model, prompt, or configuration changes?

Important requirement, weight 2

08 Are safeguards tested for high-risk outputs (dangerous, violent, or CBRN-related content)?

Critical requirement, weight 3

03

Confabulation and Output Quality

09 Are measures in place to reduce and detect confabulated (hallucinated) outputs, e.g., grounding or retrieval?

Critical requirement, weight 3

10 Are factual accuracy and output quality measured for key GenAI use cases?

Important requirement, weight 2

11 Are users clearly warned that GenAI outputs may be inaccurate and require verification?

Important requirement, weight 2

04

Content Provenance and Information Integrity

12 Is AI-generated content labelled or disclosed to users where appropriate?

Critical requirement, weight 3

13 Are provenance techniques (watermarking, metadata, C2PA-style credentials) used or planned for generated content?

Important requirement, weight 2

14 Are controls in place to prevent your GenAI systems being used for disinformation or impersonation?

Important requirement, weight 2

05

Data Privacy and Intellectual Property

15 Are controls in place to prevent sensitive or personal data leaking into prompts, training data, or outputs?

Critical requirement, weight 3

16 Are retention and usage terms of GenAI vendors reviewed (e.g., whether your data is used for training)?

Critical requirement, weight 3

17 Are intellectual property risks of generated content assessed (copyright, licensing of training data and outputs)?

Important requirement, weight 2

18 Is there guidance on what data classes may and may not be used with GenAI tools?

Critical requirement, weight 3

06

Harmful Content and Bias

19 Are content filters or moderation controls in place for obscene, degrading, or abusive outputs?

Critical requirement, weight 3

20 Is harmful bias in GenAI outputs assessed and mitigated for your use cases?

Important requirement, weight 2

21 Are there mechanisms for users to report harmful or incorrect outputs?

Important requirement, weight 2

07

Human-AI Configuration and Oversight

22 Is human review required for consequential decisions or externally published GenAI content?

Critical requirement, weight 3

23 Are roles defined for who may deploy, configure, and operate GenAI systems?

Important requirement, weight 2

24 Are measures in place to prevent over-reliance on GenAI (automation bias) in critical workflows?

Important requirement, weight 2

08

Value Chain and Third Parties

25 Are third-party GenAI models, APIs, and plugins risk-assessed before integration?

Critical requirement, weight 3

26 Are contractual and security requirements defined for GenAI vendors (SLAs, incident notification, data handling)?

Important requirement, weight 2

27 Do you track model/version changes by vendors and re-test when they occur?

Important requirement, weight 2

09

Monitoring, Incidents and Decommissioning

28 Are GenAI systems monitored in production (misuse patterns, drift, degraded output quality)?

Critical requirement, weight 3

29 Is there an incident response process covering GenAI-specific incidents (leakage, harmful output, jailbreak)?

Critical requirement, weight 3

30 Are processes in place to disable, roll back, or decommission GenAI features when risks exceed tolerance?

Important requirement, weight 2

Results are generated instantly. Self-assessment only: it is not legal advice, an audit opinion or certification, and completing it confers no conformity status.