Why is global AI governance so difficult to coordinate?

Multiple Overlapping Obstacles, Not One Single Barrier

Global AI governance is widely recognized as genuinely difficult to coordinate, and this difficulty stems from several distinct but interacting factors rather than any single obstacle. Countries differ significantly in their economic incentives around AI, their national security concerns, their existing legal and regulatory traditions, and their current level of AI development and capability — and each of these differences makes reaching the kind of broad consensus that binding international agreements typically require considerably harder to achieve.

Recognizing these distinct factors helps explain why, despite years of increasing international dialogue on AI governance, binding global rules remain limited compared with the scope of AI’s global impact.

Economic Competition and Strategic Incentives

AI has come to be viewed by many governments as a technology of significant economic and strategic importance, closely tied to national competitiveness, productivity growth, and technological leadership. This creates a genuine incentive for countries to want to maintain flexibility in how they develop and deploy AI domestically, rather than accepting binding international constraints that might be perceived as putting their own AI industries or capabilities at a disadvantage relative to competitors. This dynamic is a familiar pattern in international governance more broadly — technologies viewed as strategically important tend to be harder to govern collectively than those viewed as more neutral or universally beneficial without significant competitive stakes attached.

National security considerations compound this further. Some AI capabilities have potential military or strategic applications, which makes governments understandably cautious about the kind of transparency and information-sharing that meaningful international cooperation on AI safety might otherwise call for.

Differing Legal Systems and Uneven Development

Countries also differ substantially in their existing legal and regulatory traditions, which shapes how comfortable and able they are to adopt binding international AI rules. Some jurisdictions favor more precautionary, rights-focused regulatory approaches, while others tend toward lighter-touch, innovation-focused frameworks — differences that predate AI but that make harmonizing binding global rules for AI specifically more complex to negotiate.

Additionally, countries currently vary considerably in their level of AI development and capability, creating asymmetries that complicate negotiations. Countries at the frontier of AI development may have different priorities and concerns than countries that are earlier in AI adoption, and reaching terms that feel fair and workable to all parties, given this uneven starting point, adds another layer of difficulty to international negotiations.

A Pace Mismatch

Finally, there’s a structural mismatch between how quickly AI technology continues to evolve and how slowly traditional international governance processes, including treaty negotiation and ratification, tend to move. By the time extensive international negotiations might conclude on a specific set of AI capabilities or risks, the technology itself may have evolved considerably, which complicates efforts to build durable, binding agreements around a rapidly moving target.

Bottom Line

Global AI governance is difficult to coordinate because of a combination of competing economic and strategic incentives, national security sensitivities, differing legal traditions, and uneven AI development across countries, all compounded by a pace of technological change that regularly outstrips the slower, more deliberate pace of traditional international treaty-making.

Go deeper

• Are Schools in Different Countries Adopting Classroom AI at the Same Pace?
• Could AI Chip Manufacturing Become a Geopolitical Flashpoint?
• What Are AI Chip Export Controls and Why Do They Exist?