Regulatory capacityâgovernmentâs ability to effectively oversee and regulate AIâis a critical constraint on AI governance. Even well-designed regulations fail if regulators canât understand AI systems, canât monitor compliance, and canât adapt rules to rapid technological change. Current regulatory capacity for AI is severely limited across most jurisdictions.
The gaps are substantial. Most regulatory agencies have few staff with AI expertise, making it difficult to evaluate technical claims or design appropriate requirements. Government AI budgets are tiny compared to industry R&D spending. The regulatory processâfrom proposed rule to implementationâtakes years, while AI capabilities advance in months. And AI oversight is fragmented across multiple agencies with overlapping and incomplete mandates.
Efforts to build capacity are underway. AI Safety Institutes have been established in the US, UK, Japan, Singapore, and other countries to develop technical expertise. The EU AI Office is staffing up to implement the AI Act. But the scale of investment remains far below what effective oversight would require, and the fundamental speed mismatch between regulation and AI development persists.