The Future of International Scientific Assessments of AI's Risks
Hadrien Pouget, Claire Dennis, Jon Bateman, Robert F. Trager, Renan Araujo, Haydn Belfield, Belinda Cleeland, Malou Estier, Gideon Futerman, Oliver Guest, Carlos Ignacio Gutierrez, Vishnu Kannan, Casey Mahoney, Matthijs Maas, Charles Martinet, Jakob Mökander, Kwan Yee Ng, Seán Ó hÉigeartaigh, Aidan Peppin, Konrad Seifert, Scott Singer, Maxime Stauffer, Caleb Withers, Marta Ziosi
Abstract
Effective international coordination to address AI’s global impacts demands a shared, scientifically rigorous understanding of AI risks. This paper examines the challenges and opportunities in establishing international scientific consensus in this domain. It analyzes current efforts, including the UK-led International Scientific Report on the Safety of Advanced AI and emerging UN initiatives, identifying key limitations and tradeoffs. The authors propose a two-track approach: 1) a UN-led process focusing on broad AI issues and engaging member states, and 2) an independent annual report specifically focused on advanced AI risks. The paper recommends careful coordination between these efforts to leverage their respective strengths while maintaining their independence. It also evaluates potential hosts for the independent report, including the network of AI Safety Institutes, the OECD, and scientific organizations like the International Science Council. The proposed framework aims to balance scientific rigor, political legitimacy, and timely action to facilitate coordinated international action on AI risks.
Cite
@techreport{2024_pouget_et_al_international_scientific_assessments,
title = {{The Future of International Scientific Assessments of AI's Risks}},
author = {Hadrien Pouget and Claire Dennis and Jon Bateman and Robert F. Trager and Renan Araujo and Haydn Belfield and Belinda Cleeland and Malou Estier and Gideon Futerman and Oliver Guest and Carlos Ignacio Gutierrez and Vishnu Kannan and Casey Mahoney and Matthijs M. Maas and Charles Martinet and Jakob Mökander and Kwan Yee Ng and Seán Ó hÉigeartaigh and Aidan Peppin and Konrad Seifert and Scott Singer and Maxime Stauffer and Caleb Withers and Marta Ziosi},
year = {2024},
institution = {Oxford Martin AI Governance Initiative and Carnegie Endowment for International Peace},
}