ai-safe2-framework

AISM Self-Assessment Tool

AISM Self-Assessment Tool

Documentation updated: 2026-10-02. Clarifies the existing Sovereignty Score and introduces TCP and Evidence Assurance v1.0 as separate assessment methods. See the assessment guide for the three instruments, Technology Card workflow, and worked example. Original methodology dates below identify their baseline; this update does not change scoring formulas.

Using technology evidence in an organizational assessment

Record candidate evidence in the Technology Card before deciding which organizational assessment cells it supports. Confirm local deployment, ownership, control coverage, tested boundaries, and complementary controls. A paper, released code, or vendor test cannot fill an organizational cell by itself. Keep assurance v1.0 separate from current CLI grades; unreviewed cells remain unscored rather than inheriting a product contribution.

AI Sovereign Maturity (AISM) Model: Organizational Assessment Questionnaire

Version: 1.0
Date: March 2026
Organization: Cyber Strategy Institute
Framework: AI SAFE2 v3.1 — AISM Scoring Methodology Purpose: Self-assessment for organizations to determine their AISM maturity level


Instructions

How to Use This Assessment

  1. Assemble your assessment team — Include representatives from: Security/CISO office, AI/ML Engineering, Legal/Compliance, IT Operations, and Executive Leadership
  2. For each checklist item, mark one of:
    • ✅ Implemented — Control is fully operational with evidence
    • 🔶 Partial — Control exists but is incomplete or inconsistent
    • ❌ Not Implemented — Control does not exist
    • ⬜ N/A — Not applicable to your organization
  3. Rate each section using the three AISM metrics (Coverage, Robustness, Sovereignty Assurance) as Low/Medium/High
  4. Calculate your score using the AISM Scoring Matrix (see Scoring Methodology document)
  5. Document evidence for each implemented control

Scoring Quick Reference

Score Metrics (Coverage / Robustness / Sovereignty) Maturity Level
5 High / High / High Sovereignty (alias: Autonomous Governance)
4 High / High / Medium Control (alias: Sovereign)
3 Medium combinations (HMM, HHL, MMM, etc.) Governance (alias: Controlled)
2 Low-medium combinations (MML, MLL, HLL) Visibility (alias: Aware)
1 Low / Low / Low Chaos (alias: Reactive)

Pillar 1: Sanitize / Isolate (P1)

Topic 1: Sanitize — Input Validation, Data Filtering, Cleansing

Level 1 (Chaos; alias: Reactive) — Baseline Controls

Level 2 (Visibility; alias: Aware) — Foundation Controls

Level 3 (Governance; alias: Controlled) — Formalized Controls

Level 4 (Control; alias: Sovereign) — Advanced Controls

Level 5 (Sovereignty; alias: Autonomous Governance) — Optimizing Controls

Section Rating:


Topic 2: Isolate — Containment, Sandboxing, Boundary Enforcement

Level 1 (Chaos; alias: Reactive)

Level 2 (Visibility; alias: Aware)

Level 3 (Governance; alias: Controlled)

Level 4 (Control; alias: Sovereign)

Level 5 (Sovereignty; alias: Autonomous Governance)

Section Rating:


Pillar 2: Audit / Inventory (P2)

Topic 3: Audit — Verification, Accountability, Tracking

Level 1 (Chaos; alias: Reactive)

Level 2 (Visibility; alias: Aware)

Level 3 (Governance; alias: Controlled)

Level 4 (Control; alias: Sovereign)

Level 5 (Sovereignty; alias: Autonomous Governance)

Section Rating:


Topic 4: Inventory — Asset Mapping, Dependencies, Documentation

Level 1 (Chaos; alias: Reactive)

Level 2 (Visibility; alias: Aware)

Level 3 (Governance; alias: Controlled)

Level 4 (Control; alias: Sovereign)

Level 5 (Sovereignty; alias: Autonomous Governance)

Section Rating:


Pillar 3: Fail-Safe / Recovery (P3)

Topic 5: Fail-Safe — Shutdowns, Error Handling, Resilience

Level 1 (Chaos; alias: Reactive)

Level 2 (Visibility; alias: Aware)

Level 3 (Governance; alias: Controlled)

Level 4 (Control; alias: Sovereign)

Level 5 (Sovereignty; alias: Autonomous Governance)

Section Rating:


Topic 6: Recovery — Backups, Restoration, Continuity

Level 1 (Chaos; alias: Reactive)

Level 2 (Visibility; alias: Aware)

Level 3 (Governance; alias: Controlled)

Level 4 (Control; alias: Sovereign)

Level 5 (Sovereignty; alias: Autonomous Governance)

Section Rating:


Pillar 4: Engage / Monitor (P4)

Topic 7: Engage — Human Oversight, Intervention, Interaction

Level 1 (Chaos; alias: Reactive)

Level 2 (Visibility; alias: Aware)

Level 3 (Governance; alias: Controlled)

Level 4 (Control; alias: Sovereign)

Level 5 (Sovereignty; alias: Autonomous Governance)

Section Rating:


Topic 8: Monitor — Observation, Anomaly Detection, Logging

Level 1 (Chaos; alias: Reactive)

Level 2 (Visibility; alias: Aware)

Level 3 (Governance; alias: Controlled)

Level 4 (Control; alias: Sovereign)

Level 5 (Sovereignty; alias: Autonomous Governance)

Section Rating:


Pillar 5: Evolve / Educate (P5)

Topic 9: Evolve — Threat Adaptation, Continuous Improvement

Level 1 (Chaos; alias: Reactive)

Level 2 (Visibility; alias: Aware)

Level 3 (Governance; alias: Controlled)

Level 4 (Control; alias: Sovereign)

Level 5 (Sovereignty; alias: Autonomous Governance)

Section Rating:


Topic 10: Educate — Training, Culture, Awareness

Level 1 (Chaos; alias: Reactive)

Level 2 (Visibility; alias: Aware)

Level 3 (Governance; alias: Controlled)

Level 4 (Control; alias: Sovereign)

Level 5 (Sovereignty; alias: Autonomous Governance)

Section Rating:


AI SAFE² v3.1 Protocol Governance Overlay

Complete this overlay for any deployment that exposes MCP or equivalent agent-to-tool interfaces. It is a deployment-readiness check and does not add a sixth pillar or change the 30-cell AISM scoring formula.

Overlay decision: ☐ Ready ☐ Ready with restrictions ☐ Hold

Evidence references: ________________


Overall AISM Score Summary

Pillar Topic Score (1–5)
P1: Sanitize / Isolate T1: Sanitize ___/5
  T2: Isolate ___/5
  Pillar P1 Average ___/5
P2: Audit / Inventory T3: Audit ___/5
  T4: Inventory ___/5
  Pillar P2 Average ___/5
P3: Fail-Safe / Recovery T5: Fail-Safe ___/5
  T6: Recovery ___/5
  Pillar P3 Average ___/5
P4: Engage / Monitor T7: Engage ___/5
  T8: Monitor ___/5
  Pillar P4 Average ___/5
P5: Evolve / Educate T9: Evolve ___/5
  T10: Educate ___/5
  Pillar P5 Average ___/5
     
Overall AISM Sovereignty Score   ___/5

Maturity Classification

Score Level Classification
4.50 – 5.00 5 Sovereignty (alias: Autonomous Governance) — AI sovereignty fully realized
3.50 – 4.49 4 Control (alias: Sovereign) — Runtime control and adaptive oversight
2.50 – 3.49 3 Governance (alias: Controlled) — Formalized controls and documented processes
1.50 – 2.49 2 Visibility (alias: Aware) — Basic controls with inconsistent application
1.00 – 1.49 1 Chaos (alias: Reactive) — Ad hoc, incident-driven governance

The canonical names are defined in maturity-model.md. Aliases are retained only for compatibility with earlier assessment language.


Assessment Completion

Item Details
Organization Name  
Assessment Date  
Assessment Team Lead  
Team Members  
Overall AISM Score  
Maturity Classification  
Priority Gaps Identified  
Next Review Date  

References

  1. Cyber Strategy Institute (2026). AI SAFE2 Framework v3.1.
  2. Dotan et al. (2024). IEEE/NIST AI RMF Flexible Maturity Model.
  3. Microsoft Research (2023). Responsible AI Maturity Model.
  4. Cloud Security Alliance (2025). AI Controls Matrix (AICM) v1.0.
  5. Bernardo et al. (2025). NIST CSF Maturity Assessment Framework.
  6. Darling et al. (2026). Maturity-Based Certification for Embodied AI.

© 2026 Cyber Strategy Institute. Licensed under CC BY 4.0.