The Legal Priorities Project has published a new report (link, PDF, SSRN) surveying models, different examples, and proposals for international institutions for AI governance.
This literature review examines a range of institutional models that have been proposed over the year for the international governance of AI. The review specifically focuses on proposals that would involve the creation of new international institutions for AI. As such, it focuses on seven models for international AI institutions with distinct functions. These models are:
Scientific consensus building
Political consensus-building and norm-setting
Coordination of policy and regulation
Enforcement of standards or restrictions
Stabilization and emergency response
International joint research
Distribution of benefits and access
Part I consists of the literature review. For each model, we provide (i) a description of each model’s functions and types; (ii) the most common examples of each model; (iii) some under-explored examples that are not (often) mentioned in the AI governance literature but that show promise; (iv) a review of proposals for the application of that model to the international regulation of AI; and (v) critiques of the model both generally and in its potential application to AI.
Part II briefly discusses some considerations for further research concerning the design of international institutions for AI, including the effectiveness of each model at accomplishing its aims; treaty-based regulatory frameworks; other institutional models not covered in this review; the compatibility of institutional functions; and institutional options to host a new international AI governance body.
Overall, the review covers seven institutional models, as well as more than thirty-three commonly invoked examples of those models, twenty-two additional examples, and forty-seven proposals of new AI institutions based on those models. Table 1 summarizes these findings.
Table 1: Overview of institutional models, examples, and proposed institutions surveyed
Model
Common examples
Under- explored examples
Proposed AI institutions
Scientific consensus- building
IPCC
IPBES
SAP
CEP
WMO
IPAI
Commission on Frontier AI
Intergovernmental Panel on Information Technology
Political consensus- building and norm-setting
COPs (e.g. UNFCCC COP)
OECD
G20
G7
ISO
IEC
ITU
Various soft law instruments
Lysøen Declaration
Codex Alimentarius Commission
BRICS
IAIO
Emerging Technology Coalition
IAAI
Data Governance Structure
Data Stewardship Organization
International Academy for AI Law and Regulation
Coord. of policy and regulation
WTO
ICAO
IMO
IAEA
FATF
UNEP
ILO
UNESCO
EMEP
World Bank
IMF
WSIS
Advanced AI Governance Organisation
IAIO
EU AI Agency
GAIA
Generative AI global governance body
Coordinator and Catalyser of International AI Law
Enforcement of standards or restrictions
IAEA (Department of Safeguards)
Nuclear Suppliers Group
Wassenaar Arrangement
Missile Technology Control Regime
Open Skies Consultative Commission
Atomic Development Authority
OPCW
BWC Implementation Unit
IMO
CITES Secretariat
UN AI control agency
Global watchdog agency
International Enforcement Agency
Emerging Technologies Treaty
IAIA (multiple)
UN Framework Convention on AI (UNFCAI) & Protocol on AI, supported by Intergovernmental Panel on AI, AI GLobal Authority, and supervisory body
Advanced AI Governance Organization
AIEA for Superintelligence
NPT+
Multilateral AI governance initiative
International AI Safety Agency
Advanced AI chips registry
Code of conduct for state behavior
AI CBMs
Open Skies for AI
Bilateral US-China regime
Stabilization and emergency response
FSB
UNDRR
WHO
IAEA
Global Foresight Observatory
Geotechnology Stability Board
Intern. joint research
CERN
ITER
ISS
Human Genome Project
Atomic Development Authority (proposed)
James Webb Telescope
LIGO
AI Safety Project
Clearinghouse for research into AI
Benevolent AGI Treaty
Multilateral Artificial Intelligence Research Institute (MAIRI)
Neutral hub for AI research
UN AI Research Organization (UNAIRO)
CERN for AI
International supercomputing research facility
Joint international AI project
Multilateral AGI Consortium
European Artificial Intelligence megaproject
Distribution of benefits and access
Gavi
Vaccine Alliance
Global Fund to Fight AIDS, Tuberculosis and Malaria
This post presents the executive summary from Giving What We Can’s impact evaluation for 2025. At the end of this post we share links to more information, including the full report and...
I used AI to fix transcription errors, rerrarange the ideas, and suggest tweaks to the title and some sentences.
Three of the most exciting projects to come out of EA in recent years are, in a vague sense, CEA spinouts:
* Kairos is directly a spinout of CEA and now handles most support for university AI safety groups. Basically everyone I've found who knows them is really excited about what they do
* NEST is an opinionated ideas-fi...
