Standards & Frameworks
The landscape
No single framework covers everything. Most organisations face overlapping requirements from multiple standards — and the gaps between them are where problems hide.
IQ:NS models 50+ frameworks and domain models from ./ontologies/v1/ as structured ontologies. Here’s what they cover.
IQ:NS does not prioritise a single set of regulations; it cherry-picks the relevant ontology models for each context from the repository.
Horizontal Frameworks
EU AI Act
- Applies to: Any organisation offering or deploying AI in the EU
- Covers: Risk classification, conformity assessment, transparency, technical documentation, human oversight
- Enforcement: Fines up to 7% of global revenue
NIST AI Risk Management Framework
- Applies to: Voluntary, but increasingly referenced by US regulators and procurement
- Covers: Four core functions — Govern, Map, Measure, Manage
- Why it matters: The most operationally actionable framework available
ISO 42001
- Applies to: Organisations seeking formal AI management certification
- Covers: Plan-do-check-act management system, Annex A controls, internal audit
EU GDPR (AI provisions)
- Applies to: Any organisation processing EU resident personal data with AI
- Covers: Article 22 (automated decisions), profiling transparency, data minimization
Sector-Specific Standards
Financial Services
| Framework | Jurisdiction | Focus |
|---|---|---|
| SR 11-7 | US | Model risk management |
| EBA AI Guidelines | EU | Credit institution AI oversight |
| MAS AI Governance | Singapore | Financial institution AI governance |
Healthcare
| Framework | Jurisdiction | Focus |
|---|---|---|
| FDA AI/ML SaMD | US | Clinical AI and software as medical device |
| ONC Trustworthy AI | US | Health IT AI standards |
Security
| Framework | Focus |
|---|---|
| OWASP LLM Top 10 | LLM security vulnerabilities |
| MITRE ATLAS | Adversarial AI threat landscape |
| Google SAIF | Secure AI framework |
Ethics & Principles
| Framework | Origin |
|---|---|
| OECD AI Principles | OECD |
| UNESCO AI Ethics | UNESCO |
| IEEE 7000 | IEEE |
How frameworks overlap
Different terminology, same concepts:
| Concept | EU AI Act | NIST AI RMF | ISO 42001 |
|---|---|---|---|
| Risk management | Article 9 | Map function | Clause 6.1 |
| Human oversight | Article 14 | Govern 1.4 | A.8.4 |
| Technical documentation | Article 11 | Map 3.4 | A.6.2 |
| Bias and fairness | Article 10(2)(f) | Measure 2.6 | A.10.3 |
| Transparency | Chapter IV | Govern 4.1 | A.8.3 |
IQ:NS maps these crosswalks structurally — so you can see where one concept spans multiple standards and where genuine gaps exist.