Proposed Model AI Governance Framework for Generative AI

• Authority: Government of Singapore
• Jurisdiction: Other countries
• Collections: Miscellaneous documents
• Most recent activity: Proposed
• Most recent activity date: 2024-01-16

Risk subdomain coverage

Good coverage — Comprehensive governance measures or mitigations with clear procedures relating to the risk subdomain are explicitly described.

• Exposure to toxic content
• Compromise of privacy
• AI system security vulnerabilities
• False or misleading information
• Pollution of information ecosystem
• Governance failure
• AI pursuing its own goals
• AI possessing dangerous capabilities
• Lack of capability or robustness
• Lack of transparency or interpretability

Minimal coverage — Brief mention but no specific mitigations, recommendations or governance measures relating to the risk subdomain are described.

• Unfair discrimination and misrepresentation
• Disinformation, surveillance, and influence at scale
• Cyberattacks, weapon development or use
• Fraud, scams, and targeted manipulation
• Overreliance and unsafe use
• Power centralization and unfair distribution
• Increased inequality and decline in employment
• Economic and cultural devaluation of human effort
• Competitive dynamics
• Environmental harm

All other risk subdomains had no coverage in this document.

Risk Coverage Summary

The document has good coverage of approximately 12-14 subdomains, with strong focus on AI system security (2.2), false information/misinformation (3.1, 3.2), malicious actors (4.1, 4.2, 4.3), overreliance (5.1), competitive dynamics (6.4), governance failure (6.5), and AI safety failures (7.1, 7.2, 7.3, 7.4). Coverage is concentrated in security, misuse prevention, misinformation, and AI system safety domains.

Sector coverage

Good coverage — Comprehensive governance measures or mitigations with clear procedures relating to the sector are explicitly described.

• Information
• Scientific Research & Development Services

Minimal coverage — Brief mention but no specific mitigations, recommendations or governance measures relating to the sector are described.

• Finance & Insurance
• Professional & Technical Services
• Educational Services
• Health Care & Social Assistance
• Arts, Entertainment & Recreation
• Public Administration (excl. National Security)
• National Security

All other sectors had no coverage in this document.

Sectors Summary

This is a cross-sectoral model framework that applies broadly to generative AI development and deployment across all industries. It explicitly mentions healthcare, education, and finance as examples of sectors where AI is used, and addresses public sector AI adoption. The framework is sector-agnostic in its core governance principles but acknowledges sector-specific requirements may apply.

Legislative Status

Soft Law

Legislative Status Summary

This is a non-binding Model AI Governance Framework that provides recommendations and best practices using voluntary language throughout. It explicitly states it 'seeks to set forth a systematic and balanced approach' and welcomes feedback, indicating its advisory rather than mandatory nature.

Actors

Proposer

• Entity Types: AI_Governance_Actor
• Entities: Government of Singapore
• Summary: The document is explicitly authored by the Government of Singapore as indicated in the document information and references to Singapore's previous AI governance frameworks and initiatives.

Target

• Entity Types: AI_Developer, AI_Deployer, AI_Infrastructure_Provider
• Entities: model developers; application deployers; cloud service providers
• Summary: The framework explicitly identifies and addresses multiple actors in the AI development chain including model developers, application deployers, and cloud service providers throughout the document.

Enforcer

• Summary: The document does not specify enforcement bodies or mechanisms as it is a voluntary model framework rather than binding regulation. References to enforcement are in the context of examples from other jurisdictions or future considerations.

Monitor

• Entity Types: AI_Governance_Actor
• Entities: Information Sharing and Analysis Centres; relevant authorities; third-party testing entities; audit and professional services firms
• Summary: The document discusses monitoring through incident reporting to Information Sharing and Analysis Centres, third-party testing and assurance mechanisms, and references to government oversight for high-risk models.

AI Lifecycle Stages

Lifecycle stages covered by the documents, as defined in the OECD Framework for the Classification of AI Systems

• Plan and Design
• Collect and Process Data
• Build and Use Model
• Verify and Validate
• Deploy
• Operate and Monitor

AI Lifecycle Summary

The document comprehensively covers all stages of the AI lifecycle with particular emphasis on Build and Use Model, Verify and Validate, Deploy, and Operate and Monitor stages. It addresses planning through accountability frameworks, data collection and processing extensively, model development best practices, comprehensive evaluation and testing requirements, deployment transparency, and ongoing monitoring through incident reporting.

Technical Scope

Terms relating to technical scope that are mentioned or implicitly referenced in the document

• AI Models
• AI Systems
• Generative AI
• Open Weight / Open Source

Technical Scope Summary

The document explicitly focuses on generative AI throughout, with extensive discussion of AI models and AI systems. It addresses model development, deployment, and governance but does not use specific terminology like frontier AI, general purpose AI, or compute thresholds. It implicitly covers both open-weight and closed-source models through discussions of different model types.