AI Governance Needs Infrastructure, Not More Consensus

The United Nations will convene its first Global Dialogue on AI Governance this July facing an unusual challenge: widespread agreement on AI's risks paired with little consensus on solutions. But according to Katy Knight, President and CEO of the Siegel Family Endowment, the conversation is focused on the wrong problem.

The real barrier to effective AI governance isn't philosophical disagreement—it's the absence of operational infrastructure needed to coordinate action across borders, Knight argues in an essay originally published by the World Economic Forum.

The power imbalance problem

A small number of frontier technology companies currently control the technical talent, computing resources, evaluation capabilities, and narratives around AI development. This creates a fundamental power imbalance: governments and philanthropic funders make high-stakes decisions based largely on information provided by the very actors they're attempting to regulate.

Meanwhile, the institutions that should be coordinating—national governments, research organizations, civil society groups, and philanthropic foundations—operate in disconnected silos. Governments pursue parallel strategies without systematic knowledge sharing. Researchers work across fragmented networks. Civil society organizations struggle to access technical expertise.

The problem is particularly acute for countries in the Global South, which remain systematically excluded from designing frameworks that will shape their technological futures. "We find ourselves in a predicament of inaction not due to a lack of effort, but a lack of connective tissue," Knight writes.

What governance infrastructure looks like

Knight outlines five critical components for effective AI governance infrastructure:

Durable peer-learning networks that allow countries facing similar policy challenges to share experiments, failures, and lessons in real time rather than waiting for annual convenings.

Rapid knowledge exchange mechanisms that help decision-makers understand implications of significant developments—whether breakthrough capabilities, safety incidents, or innovative regulatory approaches—and coordinate responses quickly.

Practice documentation systems that capture, evaluate, and openly share governance experiments happening worldwide, making learning cumulative rather than fragmented.

Independent technical evaluation capacity that enables governments to assess systems, test claims, and understand technical realities beneath policy frameworks without relying solely on industry sources.

Space for independent voices from workers, educators, artists, Indigenous communities, and others whose futures will be shaped by AI—ensuring governance reflects lived realities rather than abstract assumptions.

Why it matters

The infrastructure gap in AI governance represents a structural power imbalance that undermines democratic oversight. When a handful of companies control both the technology and the expertise needed to evaluate it, even well-designed regulations become difficult to implement. Building independent evaluation capacity and coordination mechanisms isn't just about better policy—it's about whether governments can meaningfully govern transformative technology at all.

Distributed responsibility

No single institution can build this infrastructure alone, Knight notes. Governments must establish rules and coordinate internationally. Researchers can translate technical developments into actionable insights. Civil society ensures governance reflects real-world impacts. Industry contributes expertise to standards development.

Philanthropy has a unique role funding the institutions that don't fit market incentives or government mandates—independent evaluation capabilities, technical capacity-building programs, and cross-sector coordination mechanisms.

Knight argues the UN AI Dialogue's success shouldn't be measured by the strength of its final declaration, but by whether it creates durable mechanisms for learning, coordination, and collective action. "If AI governance is ultimately an infrastructure challenge, then Geneva should be remembered not as the place where the world reached consensus, but as the place where it began building the capacity to act on it."

These details were first reported by the World Economic Forum.