Skip to main content

​
Dimension Framework

Each dimension measures a distinct aspect of AI governance. Together, they provide a comprehensive view of how a company governs artificial intelligence.

​
D1: Leadership & Accountability

Measures: Board oversight structure, C-suite AI ownership, named responsibility chains, governance team composition, strategic alignment, safety culture. Committee: Full Board | Standards: ISO 42001 Clause 5, NIST GOVERN 1.1-1.7, EU AI Act Article 4 Key indicators: Named C-suite AI accountability, board AI competency and composition, board agenda integration, committee charter language, safety culture signals.

​
D2: AI Safety & Robustness

Measures: Testing protocols, red-teaming, deployment safeguards, incident response, responsible scaling, governance under stress. Committee: Risk | Standards: ISO 42001 Annex B.6, NIST MEASURE 2.5-2.13, EU AI Act Articles 9, 15 Key indicators: Pre-deployment safety testing, staged rollout protocols, responsible scaling policies, dual-use testing, agentic AI governance, defence-in-depth architecture. Highest-value signal: Governance Under Stress (2I), which measures documented instances where governance constrained commercial decisions.

​
D3: Transparency & Disclosure

Measures: Public reporting on AI systems, model documentation, risk factor disclosure, stakeholder communication. Committee: Audit | Standards: EU AI Act Article 13, ISO 42001 Annex B.7, NIST GOVERN 4 Key indicators: AI system inventory disclosure, model card or system card publication, risk factor specificity in SEC filings, algorithmic impact assessments.

​
D4: Data Ethics & Privacy

Measures: Training data governance, consent frameworks, privacy-by-design, data quality controls, cross-border compliance. Committee: Risk | Standards: EU AI Act Article 10, NIST MEASURE 2.6, ISO 42001 Annex B.8 Key indicators: Training data documentation, consent and licensing frameworks, privacy impact assessments, data quality and bias testing protocols.

​
D5: Governance Infrastructure

Measures: Policies, processes, organizational architecture, internal audit, change management, resource allocation. Committee: Nominating/Governance | Standards: ISO 42001 Clauses 6-10, NIST GOVERN 1.2-1.5 Key indicators: AI governance policy existence and specificity, internal audit of AI systems, change management protocols, dedicated governance resources, model risk management framework.

​
D6: Fairness & Bias Mitigation

Measures: Bias testing, demographic impact analysis, fairness metrics, remediation processes, accessibility. Committee: Compensation/HR | Standards: EU AI Act Article 10(2), NIST MEASURE 2.6 Key indicators: Bias testing methodology, demographic impact reporting, fairness metric definition and monitoring, third-party bias audits, remediation evidence.

​
D7: Societal, Workforce & Environmental Impact

Measures: Workforce transition planning, environmental footprint of AI compute, community impact, responsible deployment in sensitive contexts. Committee: Compensation/HR | Standards: EU AI Act Article 14.4, NIST GOVERN 6.1-6.2 Key indicators: Workforce displacement planning, reskilling investment, compute carbon footprint reporting, community engagement on AI deployment, responsible use policies.

​
D8: Regulatory & Compliance Posture

Measures: Regulatory preparedness, compliance program maturity, cross-jurisdictional alignment, proactive engagement with regulators. Committee: Audit | Standards: EU AI Act (full compliance), NIST AI RMF GOVERN Key indicators: Regulatory mapping and gap analysis, compliance team composition, cross-jurisdictional preparedness, regulatory engagement (comment letters, consultations), conformity assessment readiness.

​
D9: Cybersecurity & AI Resilience

Measures: AI-specific security controls, adversarial robustness, supply chain security, incident detection and response. Committee: Risk | Standards: EU AI Act Article 15, NIST MEASURE 2.7 Key indicators: AI-specific threat modeling, adversarial attack testing, model integrity monitoring, supply chain security for AI components, AI incident detection capability.

​
D10: Third-Party & Supply Chain Governance

Measures: Vendor AI governance requirements, supply chain due diligence, contractual controls, monitoring of third-party AI systems. Committee: Audit | Standards: EU AI Act Articles 25-27 Key indicators: Vendor AI governance assessment, contractual AI governance requirements, third-party model evaluation, supply chain transparency, ongoing vendor monitoring.

​
D11: Stakeholder Impact Layer

Measures: External harm accountability: effects on workers, communities, environment, and other stakeholders beyond the company’s direct operations. Committee: Full Board | Standards: EU AI Act Article 14.4, NIST GOVERN 6.1-6.2 Key indicators: Stakeholder impact assessments, workforce transition evidence, environmental footprint transparency, community engagement on AI deployment, human rights due diligence. Added in The Alpha Standard v3.

​
D12: Corrigibility Assessment

Measures: Demonstrated ability to override, shut down, or correct AI systems. Evidence of governance constraining capability deployment. Committee: Risk | Standards: EU AI Act Article 14 (Human Oversight), NIST GOVERN 1.1 Key indicators: Documented shutdown or override events, governance hold on deployment, capability restriction evidence, human oversight mechanisms, corrigibility event documentation. Weight: 10% across all profiles. A Corrigibility Event (documented instance where governance overrode a deployment or capability decision) is the highest-weight evidence in the entire framework. Added in The Alpha Standard v4.