Governance Responses
Summary
Section titled “Summary”Governance responses address AI risk through rules, institutions, and coordination mechanisms. This section covers:
| Category | Description |
|---|---|
| Legislation | Enacted and proposed laws (EU AI Act, US Executive Order, etc.) |
| Compute Governance | Hardware-based governance (export controls, thresholds, monitoring) |
| International | Summits, treaties, and international coordination |
| Industry Self-Regulation | RSPs, voluntary commitments, industry standards |
| Effectiveness Assessment | Evaluating which governance approaches work |
Key Governance Mechanisms
Section titled “Key Governance Mechanisms”Hard Governance
Section titled “Hard Governance”- Legislation: Binding laws with enforcement
- Regulation: Agency rules and oversight
- International treaties: Binding agreements between nations
Soft Governance
Section titled “Soft Governance”- Standards: Industry norms and best practices
- Voluntary commitments: Lab pledges and RSPs
- Coordination: Summits, dialogues, information sharing
Major Developments (2023-2024)
Section titled “Major Developments (2023-2024)”| Development | Type | Status |
|---|---|---|
| EU AI Act | Legislation | Enacted (2024) |
| US Executive Order on AI | Executive action | Active |
| UK AI Safety Summit | International | Completed (Nov 2023) |
| Seoul AI Summit | International | Completed (May 2024) |
| Lab RSPs | Voluntary | Anthropic, OpenAI, DeepMind |
Relationship to Other Responses
Section titled “Relationship to Other Responses”Governance enables and constrains other responses:
- Technical research — Governance can fund or require safety research
- Institutions — Governance creates AISIs and oversight bodies
- Coordination — International governance enables global coordination