Shavit et al. (2023)
paperAuthors
Credibility Rating
Good quality. Reputable source with community review or editorial standards, but less rigorous than peer-reviewed venues.
Rating inherited from publication venue: arXiv
An influential policy paper proposing concrete international governance structures for advanced AI; frequently cited in discussions of multilateral AI regulation and global coordination efforts.
Paper Details
Metadata
Abstract
International institutions may have an important role to play in ensuring advanced AI systems benefit humanity. International collaborations can unlock AI's ability to further sustainable development, and coordination of regulatory efforts can reduce obstacles to innovation and the spread of benefits. Conversely, the potential dangerous capabilities of powerful and general-purpose AI systems create global externalities in their development and deployment, and international efforts to further responsible AI practices could help manage the risks they pose. This paper identifies a set of governance functions that could be performed at an international level to address these challenges, ranging from supporting access to frontier AI systems to setting international safety standards. It groups these functions into four institutional models that exhibit internal synergies and have precedents in existing organizations: 1) a Commission on Frontier AI that facilitates expert consensus on opportunities and risks from advanced AI, 2) an Advanced AI Governance Organization that sets international standards to manage global threats from advanced models, supports their implementation, and possibly monitors compliance with a future governance regime, 3) a Frontier AI Collaborative that promotes access to cutting-edge AI, and 4) an AI Safety Project that brings together leading researchers and engineers to further AI safety research. We explore the utility of these models and identify open questions about their viability.
Summary
This paper proposes four complementary international institutional models for governing advanced AI: a Commission on Frontier AI for expert consensus, an Advanced AI Governance Organization for safety standards, a Frontier AI Collaborative for equitable access, and an AI Safety Project for coordinated research. The framework draws on precedents from existing international organizations and aims to balance AI's benefits against global risks from powerful systems.
Key Points
- •Identifies four international institutional models with internal synergies: expert consensus body, standards/monitoring org, access collaborative, and safety research project.
- •Argues international coordination is necessary to manage global externalities from powerful, general-purpose AI systems that no single nation can govern alone.
- •Draws on precedents from existing international organizations (e.g., IAEA, CERN) to ground proposed governance structures in proven institutional designs.
- •Addresses both equity concerns (ensuring broad access to frontier AI) and safety concerns (setting and monitoring international safety standards).
- •Explores open questions about viability, including political feasibility, enforcement mechanisms, and how institutions would handle rapidly advancing capabilities.
Cited by 1 page
| Page | Type | Quality |
|---|---|---|
| AI Proliferation Risk Model | Analysis | 65.0 |
Cached Content Preview
# International Institutions for Advanced AI
Lewis Ho1,
Joslyn Barnhart1,
Robert Trager2,
Yoshua Bengio3,Miles Brundage4,Allison Carnegie5,Rumman Chowdhury6,
Allan Dafoe1,Gillian Hadfield7,Margaret Levi8,Duncan Snidal9
1Google DeepMind, 2Blavatnik School of Government, University of Oxford and
Centre for the Governance of AI, 3Université de Montréal and Mila, CIFAR Fellow,
4OpenAI,
5Columbia University,
6Harvard Berkman Klein,
7University of Toronto,
Vector Institute and OpenAI (independent contractor),
8Stanford University,
9Nuffield
College, University of Oxford
###### Abstract
International institutions may have an important role to play in
ensuring advanced AI systems benefit humanity. International
collaborations can unlock AI’s ability to further sustainable
development, and coordination of regulatory efforts can reduce obstacles
to innovation and the spread of benefits. Conversely, the potential
dangerous capabilities of powerful and general-purpose AI systems
create global externalities in their development and deployment, and
international efforts to further responsible AI practices could help
manage the risks they pose. This paper identifies a set of governance
functions that could be performed at an international level to address
these challenges, ranging from supporting access to frontier AI systems
to setting international safety standards. It groups these functions
into four institutional models that exhibit internal synergies and have
precedents in existing organizations: 1) a Commission on Frontier AI
that facilitates expert consensus on opportunities and risks from
advanced AI, 2) an Advanced AI Governance Organization that sets
international standards to manage global threats from advanced models,
supports their implementation, and possibly monitors compliance with a
future governance regime, 3) a Frontier AI Collaborative that promotes
access to cutting-edge AI, and 4) an AI Safety Project that brings
together leading researchers and engineers to further AI
safety research. We explore the utility of these models and identify
open questions about their viability.
## Executive Summary
Recent advances in AI have highlighted the potentially transformative
impacts of advanced systems.111By “advanced AI systems” we mean systems that are
highly capable and general purpose.
International institutions may have an
important role to play in ensuring these are globally beneficial.
International collaborations could be important for unlocking AI’s
abilities to further sustainable development and benefit humanity. Many
societies that could most benefit may not have the resources,
infrastructure or training to take advantage of current cutting-edge AI
systems. Frontier AI development may not focus on global needs, and the
economic benefits of commercial AI technologies could primarily benefit
developed countries. A failure to coordinate or harmonize regulation may
also slow innovation.
Moreover, international efforts may also be
... (truncated, 75 KB total)95d12033a9f98b31 | Stable ID: NmJkMTE3OT