Within six weeks this summer, three competing visions for governing artificial intelligence took concrete shape: the U.S. State Department hosted its second Pax Silica Summit, the United Nations convened its Global Dialogue on AI Governance, and 29 countries signed an agreement in Shanghai establishing the Chinese-led World Artificial Intelligence Cooperation Organization (WAICO).
In a new report for the Center for Strategic and International Studies (CSIS), Cascade Institute fellow Christopher Collins and Aalok Mehta, director of the CSIS Wadhwani AI Center, argue that the United States is now in an institutional race to lead global AI governance — and that China’s offer to middle powers is an attractive one. Ahead of the September 24 U.S.-China talks on AI safety, they urge U.S. policymakers to pursue a focused safety agenda, treat verification as a technical program rather than a diplomatic issue, and build a domestic model of AI governance worth exporting.
Collins is a fellow with the Polycrisis Program at the Cascade Institute and a senior associate (non-resident) with the Wadhwani AI Center at CSIS.
