AI-governance organisations · 18 catalogued
The actors who shape AI governance
Bodies that author, implement, evaluate, or influence the instruments tracked in this wiki. Grouped by role: safety institutes evaluate frontier models pre-deployment; standards bodies produce technical specs governance instruments cite; multilateral bodies coordinate across jurisdictions; frontier labs set de-facto governance via their internal policies; civil-society actors push public-comment + advocacy work.
AI Safety Institutes (5)
UK-AISI · UK
UK AI Safety Institute
Empirical evaluation of advanced AI systems for societal risks. Pre-deployment access agreements with frontier labs. Headed by Ian Hogarth (founding chair).
→ 3 instruments
US-AISI · US
US AI Safety Institute (at NIST)
AI safety evaluation, methodology development, and capacity-building hosted within NIST. Pre-deployment access agreements with frontier labs. Headed by Elizabeth Kelly.
→ 5 instruments
JP-AISI · JP
Japan AI Safety Institute
AI safety evaluation and standards co-development under the Information-technology Promotion Agency. Established Feb 2024.
→ 2 instruments
SG-AIVF · SG
Singapore AI Verify Foundation
Public-private foundation maintaining the AI Verify testing toolkit. Operates the Singapore AI Safety Institute under the AI Verify umbrella.
→ 1 instrument
EU-AI-OFFICE · EU
European AI Office
Implements the EU AI Act's general-purpose AI provisions. Designates systemic-risk models, enforces Article 55 obligations, oversees the GPAI Code of Practice.
→ 1 instrument
Standards Bodies (3)
NIST · US
National Institute of Standards and Technology
US federal non-regulatory standards agency. AI work spans the AI Risk Management Framework, GenAI Profile, the AI Safety Institute, and the Frontier AI Model Evaluation Consortium.
→ 3 instruments
ISO-IEC-JTC1-SC42 · global
ISO/IEC JTC 1/SC 42 — Artificial Intelligence
Joint Technical Committee subcommittee developing international AI standards. Published ISO/IEC 22989 (AI concepts), ISO/IEC 23894 (AI risk management), ISO/IEC 42001 (AI management systems).
→ 2 instruments
IEEE-7000-SERIES · global
IEEE Standards Association — 7000-series (AI Ethics)
Series of standards on AI ethics, transparency, and bias. IEEE 7000 (Ethically Aligned Design), IEEE 7001 (transparency), IEEE 7003 (algorithmic bias), IEEE 7010 (well-being metrics).
→ 2 instruments
Multilateral Bodies (3)
OECD-AI-WP · OECD
OECD AI Working Party (AIGO + Network of Experts)
Coordinates OECD AI Principles + OECD.AI policy observatory. Maintains the most-cited cross-national AI-policy dataset.
→ 2 instruments
GPAI · global
Global Partnership on AI
Multistakeholder initiative bridging OECD AI principles to practice. 29 member nations. Four working groups: Responsible AI, Data Governance, Future of Work, Innovation & Commercialisation.
→ 1 instrument
UN-AI-ADVISORY · UN
UN AI Advisory Body
39-member expert body advising the UN Secretary-General on international AI governance. Published 'Governing AI for Humanity' (Sep 2024) recommending a UN AI Office + global AI fund.
→ 1 instrument
Frontier AI Labs (4)
ANTHROPIC · US
Anthropic
Frontier AI developer (Claude family). Governance posture organised around the Responsible Scaling Policy (RSP), a tiered ASL framework that ties capability evaluations to deployment safeguards.
→ 3 instruments
OPENAI · US
OpenAI
Frontier AI developer (GPT family). Governance posture organised around the Preparedness Framework, a capability-tier evaluation regime; Safety & Security Committee at the board level.
→ 3 instruments
GOOGLE-DEEPMIND · US
Google DeepMind
Frontier AI developer (Gemini family). Governance posture organised around the Frontier Safety Framework (FSF), a Critical Capability Levels regime, plus internal AI Principles process.
→ 3 instruments
META-AI · US
Meta (Fundamental AI Research)
Frontier AI developer (Llama family). Distinctive open-weight governance posture: most-capable models released under permissive licenses, creating governance debate about closed-vs-open frontier-model regulation.
→ 1 instrument
Civil Society (3)
FLI · US
Future of Life Institute
Research and advocacy on catastrophic + existential AI risk. Organised the 2015 Puerto Rico conference + 2017 Asilomar Principles. Authored the Mar 2023 'Pause Giant AI Experiments' open letter (33k signatories).
→ 1 instrument
CAIS · US
Center for AI Safety
Research nonprofit focused on technical + policy AI safety. Authored the May 2023 'Statement on AI Risk' (signed by frontier-lab CEOs framing AI risk as a global priority).
→ 1 instrument
PAI · US
Partnership on AI
Multistakeholder coalition (frontier labs + civil society + academia) producing AI-deployment guidelines. Notable outputs: Synthetic Media Framework, Responsible Practices for Foundation Model Deployment.
→ 2 instruments