In his January 2026 speech at the World Economic Forum in Davos, Canadian Prime Minister Mark Carney addressed the bargain at the heart of postwar international governance: peace through prosperity. 'This bargain', he claimed, 'no longer works'. Prime Minister Carney is right, and the fix will take more than new trade deals. It requires governance built around deliberation, shared stewardship and, potentially, AI-assisted participation at scale. 

Peace through prosperity, and its limitations  

Leaders saw liberal democracy and free trade as the path to stability. The transnational institutions thus relied on rules and expanding trade to manage conflict. This approach, however, assumes problems can be managed through growth, coordinated through trade, and resolved through negotiated fixes between states. Now the problems are crossing borders faster than those strategies can handle. Climate risk, ecological degradation, and widening inequality (also known as polycrisis) spill across national borders. Governance, meanwhile, still runs through state-to-state bargaining and economic exchange.

How can global governance can manage shared planetary constraints in an interdependent world? And how can AI help? 

This raises an important question: what kind of global governance can manage shared planetary constraints in an interdependent world? And what role could large language models play in supporting that work? 

How economic growth became an end in itself

Modern international governance emerged from a specific historical context. Following the devastation of the Second World War, global leaders prioritised economic cooperation to prevent future conflict. Economic growth and trade were elevated from policy tools to 'ends in themselves'. The assumption was that expanding markets would diffuse gains broadly enough to substitute for direct political deliberation across borders.

The Bretton Woods Institutions and, later, the World Trade Organization, embodied this postwar project. The goal: stabilise national economies, expand trade, and deepen interdependence. NAFTA illustrates the model's early success. Trade among the three partners grew from about $290 billion in 1993 to more than $1.1 trillion by 2016.

​​The model is now under strain. Planetary-scale challenges impose limits that economic expansion alone cannot resolve. Polycrisis requires explicit choices about trade-offs, acceptable risk thresholds, and how costs and benefits are distributed across populations. Markets can be effective at coordinating choices through prices. But legitimacy for contested collective decisions under scarcity depends on public justification, accountable procedures, and fair bargaining.​

The global financial crisis of 2008 highlights these weaknesses. A system oriented towards short-term profit maximisation proved fragile, while governmental responses to the crisis undermined trust in domestic and international institutions. Similar patterns have since emerged across climate policy and technological governance.

Global power is becoming increasingly multipolar. This is exposing the limits of trade and economic independence as foundations for global governance

The underlying issue is not an isolated policy failure. It is a persistent reliance on economic mechanisms to perform political functions they were never designed to fulfil. What has become scarce is legitimate coordination. Electoral systems aggregate preferences intermittently within national boundaries, whereas markets aggregate willingness to pay. Neither provides sustained, inclusive deliberation across diverse populations facing shared planetary constraints. As global power becomes increasingly multipolar and political models diverge, the limits of trade and economic interdependence as foundations for global governance become more pronounced.

Deliberation works, but can it scale?

Experiments in participatory and deliberative governance offer important lessons. Citizens’ assemblies and deliberative polls show that when people have time, information, and opportunities for reasoned discussion, preferences can change and polarisation declines. Researchers and practitioners have applied these approaches to complex policy areas, including energy transitions and constitutional reform. We know that deliberation works. The challenge has been scale.  

No single intervention will address this challenge outright. But language models may offer an interesting path forward, not as decision-makers, but as infrastructure for deliberation at scale. With appropriate constraints and oversight, ​AI c​ould​ help summarise arguments across large groups, organising complex information, and mapping areas of agreement and disagreement. Contemporary work indicates that ​AI ​could ​reduce the logistical and cognitive costs of collective reasoning without displacing human judgment or political authority.

Early examples already exist. Stanford’s Online Deliberation Platform used AI-assisted moderation to structure large-scale discussions and support balanced participation across hundreds of participants in national deliberations, including recent AI governance citizen forums in Taiwan. ISWE’s Global Citizens’ Assembly has also invited citizen deliberation for global climate governance. At institutional level, similar tools could support international bodies by organising consultation processes, clarifying contested policy options, and making trade-offs more transparent to the public.

The value of AI-assisted deliberative democracy lies in enabling societies to reason through trade-offs across multiple plausible futures

​​P​lanetary governance increasingly involves choices that must be made under persistent uncertainty (also called deep uncertainity). No single forecast can be relied upon for decision-making. Legitimacy depends upon how societies reason through trade-offs across multiple plausible futures, rather than optimising for a single projected outcome.

The value of AI-assisted deliberative democracy lies in its ability to enable societies to reason together under persistent uncertainity. AI-guided tools embedded within deliberative processes ​may ​help organise large volumes of input, highlight points of agreement and disagreement, and clarify underlying assumptions. This would make collective reasoning more transparent without predetermining outcomes. 

Scale needs guardrails 

These possibilities come with clear constraints. Without careful institutional design, including public ownership standards and transparent governance of the underlying infrastructure, the same technologies risk ​reinforcing existing ​​concentrations of power or drifts toward technocratic control. Deliberative systems require transparency, accountability, and pluralism to maintain legitimacy. AI can facilitate large-scale deliberation, but it cannot determine values, resolve political conflict, or substitute for democratic judgment. Those functions remain inherently human and contestable. 

Reimagining global governance requires more than technological adoption. It calls for a shift away from treating economic expansion and surplus capital as the primary mechanisms of coordination, and toward governance models grounded in mutual responsibility and shared stewardship of planetary systems.

Governing planetary commons demands institutional systems capable of revealing points of disagreement, enabling participation, and legitimating difficult trade-offs at scale. Advances in AI offer one path towards building deliberative capacity across populations and jurisdictions previously beyond reach. Expanding that capacity at scale is one step toward aligning institutions with the realities of an interconnected and constrained world.

💊 No.15 in a Loop series on 'Rescuing Democracy'

The post 💊 Rethinking global governance with AI and deliberation  appeared first on The Loop.