AI Governance
AI governance is the discipline of setting and enforcing the rules AI systems operate under — defining who is accountable, what risks are controlled, and how trust is demonstrated to customers, regulators, investors, and the public. The central concept of this wiki.
Two axes
The founding sources stake out two complementary levels (see synthesis):
- Organizational (the firm) — management systems, inventories, lifecycle & risk management, and assurance/certification, run against standards. Exemplar: nemko-ai-governance-services by nemko-digital.
- National (the state) — regulatory regimes (israel-ai-regulation-overview) and AI strategy (israel-ai-strategy-2026) — law, policy, and strategic direction.
The two are linked by shared standards: the same frameworks (eu-ai-act, iso-iec-42001, nist-ai-rmf, OECD) a firm certifies against are the ones nations align their regimes to.
Recurring principles
- risk-based-regulation — scale oversight to impact (the principle nearly everyone shares).
- Transparency & explainability, accountability, fairness/non-discrimination, safety, privacy — the canonical ethical principles (e.g. Israel’s six in responsible-innovation).
- The live policy fork: horizontal binding law (eu-ai-act) vs sectoral soft law (responsible-innovation).
- Regulation by existing law — professional-licensing-and-ai: incumbent professional/licensing statutes (e.g. unauthorized practice of law) can restrict AI de facto, with no AI-specific act.
- Governance by technical standard — IETF work like ai-preferences and web-bot-auth writes AI access rules in protocol form (a non-state venue, targeting AI’s inputs), before — or instead of — legislation; see eff-web-under-attack-ietf.
Related
risk-based-regulation · responsible-innovation · professional-licensing-and-ai · eu-ai-act · iso-iec-42001 · nist-ai-rmf · nemko-ai-governance-services · israel-ai-regulation-overview · israel-ai-strategy-2026 · lofrayer-bar-association · ietf · ai-preferences · web-bot-auth · eff-web-under-attack-ietf