Restructuring International Organizations
New or reformed international bodies for AI governance, capable of setting standards, monitoring compliance, coordinating research, and ensuring that AI development serves global rather than narrowly national or corporate interests.
What it is:
The current international governance landscape for AI is fragmented across dozens of initiatives with overlapping mandates, limited enforcement capacity, and significant gaps in coverage. Various proposals seek to address this through institutional innovation: creating new AI-specific bodies (modeled on the IAEA, IPCC, CERN, or WTO), adapting existing organizations to take on AI governance functions, or building coordination mechanisms among the proliferating national AI safety institutes.
AI's economic impacts create specific demands on international institutions that current arrangements are not equipped to meet. Labor displacement, profit shifting, cross-border data flows, and the concentration of AI capacity in a small number of countries all require coordinated responses across domains that are currently governed by separate, poorly connected institutions: the ILO for labor standards, the WTO for trade, the OECD for taxation, the IMF and World Bank for macroeconomic stability and development. Without coordination, national responses to AI-driven economic disruption risk becoming fragmented and contradictory. International institutions can, in principle, provide the forums for countries to coordinate their responses, set shared minimum thresholds for labor protections and tax rates, and create frameworks for sharing both the gains and the adjustment costs of AI adoption across borders.
The challenge:
The fundamental obstacle is that the countries and companies with the most power in AI development have the least incentive to submit to international governance that constrains their advantage. The US, home to the world's leading AI companies, has under the Trump administration shown limited willingness to cooperate multilaterally on AI governance. Many developing countries view existing international institutions as reflecting the interests of wealthy nations, and are likely to resist governance frameworks designed without their meaningful participation. And the pace of AI development may outrun institutional design, since the years required to negotiate, establish, and staff a new international institution may exceed the window in which governance choices can meaningfully shape AI's trajectory.
Recommended Reading:
Real-world precedents:
The International Labour Organization provides a model for setting labor standards across borders, though its conventions are often weakly enforced. The ILO has adopted hundreds of conventions, protocols, and recommendations covering wages, working conditions, and collective bargaining rights.
The Financial Stability Board shows how coordination among national regulators can address cross-border economic risks without creating a supranational authority. Through its model of networked national regulators, it monitors systemic risks, issues recommendations, and conducts peer reviews of national implementation, but has no direct enforcement power — relying instead on peer review and reputational pressure.
The International Atomic Energy Agency is frequently cited as an institutional model for AI governance due to its combination of promotional and safety functions. The IAEA promotes peaceful uses of nuclear technology while administering safeguards agreements that verify compliance with nonproliferation commitments. However, critics note that nuclear technology differs fundamentally from AI in concentration, detectability, and state-centricity, limiting the analogy's applicability.