AI Safety 2 min read

AI Governance

Also known as: AI Policy, AI Regulation, AI Oversight

The frameworks, policies, standards, and oversight mechanisms that guide the development, deployment, and use of AI systems within organizations and across society.

Definition

The frameworks, policies, standards, and oversight mechanisms that guide the development, deployment, and use of AI systems within organizations and across society.

AI Safety 2 min read A

Overview

AI governance encompasses the rules, practices, and organizational structures that ensure AI systems are developed and used responsibly. It operates at multiple levels — organizational, national, and international — and addresses questions of accountability, risk management, compliance, and ethical standards.

Organizational AI Governance

  • AI Ethics Boards: Cross-functional committees that review AI projects for ethical risks
  • Model Risk Management: Frameworks for assessing and managing risks associated with AI models
  • Documentation Standards: Requirements for documenting AI systems, including model cards and datasheets
  • Monitoring and Auditing: Ongoing monitoring of AI system performance and periodic audits
  • Incident Response: Procedures for addressing AI system failures or harmful outcomes

Regulatory Frameworks

EU AI Act

The world's first comprehensive AI regulation, establishing a risk-based framework with requirements ranging from transparency obligations for low-risk systems to strict prohibitions on unacceptable uses.

U.S. Approach

A combination of executive orders, sector-specific regulations, and voluntary frameworks like the NIST AI RMF. The approach emphasizes innovation while establishing guard rails for high-risk applications.

International Coordination

Organizations like the OECD, G7, and the UN are working to coordinate AI governance globally through shared principles and interoperable standards.

Context Management and Governance

AI governance directly impacts context management practices. Data governance requirements determine what context can be provided to AI systems, data retention policies affect how long context is stored, and privacy regulations constrain how personal information can be used as context.