International AI Safety Report 2026

Singer, Scott; Munga, Jane

Research

International AI Safety Report 2026

The second International AI Safety Report is the next iteration of the comprehensive review of latest scientific research on the capabilities and risks of general-purpose AI systems. It represents the largest global collaboration on AI safety to date.

By Scott Singer and Jane Munga

Published on Feb 3, 2026

Program

Africa

The Africa Program focuses on economic, political, and transnational issues shaping Africa’s future. By conducting data-driven research, convening high-level dialogues, forging strategic partnerships, and amplifying African voices, the program addresses a crucial knowledge gap on Africa’s role in a changing global environment.

Learn More

Program

Technology and International Affairs

The Technology and International Affairs Program develops insights to address the governance challenges and large-scale risks of new technologies. Our experts identify actionable best practices and incentives for industry and government leaders on artificial intelligence, cyber threats, cloud security, countering influence operations, reducing the risk of biotechnologies, and ensuring global digital inclusion.

Learn More

Read the Publication

About the Authors

Scott Singer

Fellow, Technology and International Affairs, and Co-Director, China AI Initiative

Scott Singer is a fellow and co-director of the China AI Initiative at the Carnegie Endowment for International Peace, where his work explores the geopolitics and governance of advanced AI systems. He specializes in China’s AI ecosystem, with a particular focus on U.S.-China AI dialogue and diplomacy, U.S. and Chinese approaches to AI strategy, and comparative approaches to AI safety and risk management.

Jane Munga

Fellow, Africa Program

Jane Munga is a fellow in the Africa Program focusing on technology policy at the Carnegie Endowment for International Peace.

Technology AI

Carnegie does not take institutional positions on public policy issues; the views represented herein are those of the author(s) and do not necessarily reflect the views of Carnegie, its staff, or its trustees.

More Work from Carnegie Endowment for International Peace

Commentary
Emissary
The U.S. Has Weapons in Space. Now What?
The acknowledgment has sparked fears of a new arms race with China and Russia.
Chris O’Meara
Commentary
Emissary
Trump Had the Chance to Address U.S.-China AI Safety. He Threw Xi a Party Instead.
The summit may be remembered as a fatal blunder, when Washington, Beijing, and the AI industry collectively chose ceremony over action.
Jeffrey Prescott
Paper
AI Breakout Capacity: A Middle-Power Path Between Racing and Dependence
Germany, India, and other middle powers can’t build frontier AI alone, and they risk being cut off if they depend on the United States or China. The way through is to build the capacity for a frontier sprint while hoping never to use it.
Anton Leicht, Sam Winter-Levy
Research
The Disruptive Politics of AI
AI technology has proliferated extensively and offers authoritarian leaders three crucial advantages: scale, speed, and efficiency. As a result, it has undoubtedly become a crucial tool of repression. But AI tools also give individuals and activists the means to organize, resist, and hold their governments accountable. Today, it is far more difficult for autocratic leaders to hide their corrupt practices and govern in darkness.
Steve Feldstein, Shreya Joshi
Commentary
Emissary
Who Gets to Put Guardrails on AI?
Why governance of rapidly evolving technology will be distributed and complex.
Stewart Patrick, Scott Singer

More Work from Carnegie Endowment for International Peace

Commentary

Emissary

The U.S. Has Weapons in Space. Now What?

The acknowledgment has sparked fears of a new arms race with China and Russia.

Chris O’Meara

Trump holding a glass of champagne up in a toast, with a table of guests in the foreground

Commentary

Emissary

Trump Had the Chance to Address U.S.-China AI Safety. He Threw Xi a Party Instead.

The summit may be remembered as a fatal blunder, when Washington, Beijing, and the AI industry collectively chose ceremony over action.

Jeffrey Prescott

Two men sit on red chairs in front of a window with a NYC city skyline.

Paper

AI Breakout Capacity: A Middle-Power Path Between Racing and Dependence

Germany, India, and other middle powers can’t build frontier AI alone, and they risk being cut off if they depend on the United States or China. The way through is to build the capacity for a frontier sprint while hoping never to use it.

Anton Leicht, Sam Winter-Levy

Research

The Disruptive Politics of AI

AI technology has proliferated extensively and offers authoritarian leaders three crucial advantages: scale, speed, and efficiency. As a result, it has undoubtedly become a crucial tool of repression. But AI tools also give individuals and activists the means to organize, resist, and hold their governments accountable. Today, it is far more difficult for autocratic leaders to hide their corrupt practices and govern in darkness.

Steve Feldstein, Shreya Joshi

Commentary

Emissary

Who Gets to Put Guardrails on AI?

Why governance of rapidly evolving technology will be distributed and complex.

Stewart Patrick, Scott Singer