A Growing Global AI Divide
Artificial intelligence is developing unevenly across the world, creating a widening gap between technological powers and less-developed countries. Advanced semiconductors, computing infrastructure, and AI capabilities remain concentrated mainly in the United States and China. As a result, many countries depend on technologies they neither develop nor fully control. This dependence limits their ability to influence the standards and governance systems shaping AI’s future.
At the same time, countries outside the main technological powers face difficult choices between competing AI ecosystems. They may rely on proprietary American systems or open-weight Chinese models, while confronting concerns about safety and technical capacity. This imbalance therefore extends beyond access to technology. It also affects who has the ability to participate meaningfully in decisions about AI governance.
Against this backdrop, the United Nations has proposed a “Dialogue of Dialogues” to connect existing governance initiatives. The approach would promote technical interoperability, shared safety standards, and stronger institutional capacity among participating countries. Although it cannot challenge the technological dominance of the United States and China, it could strengthen the collective position of countries outside that rivalry.
Governing AI Through Cooperation
The need for coordination becomes clearer as AI-related risks continue to expand. These include difficulties evaluating autonomous systems, algorithmic bias, deepfake-enabled sexual violence, and the environmental costs of large-scale infrastructure. Meanwhile, global AI governance remains fragmented, with more than forty different types of governance instruments operating worldwide. Much of this framework remains voluntary and concentrated within the private sector.
To address this fragmentation, the proposed dialogue would connect governments, researchers, companies, and civil society around shared standards. Smaller regional and international coalitions could develop common approaches to issues such as deepfake detection and AI-generated sexual exploitation. Shared standards could also reduce regulatory duplication, particularly for companies operating across different jurisdictions.
Furthermore, countries could create coalitions that combine computing resources, infrastructure, and technical expertise. These arrangements would reduce dependence on the dominant AI powers while strengthening local technological capabilities. Examples include European and Brazilian initiatives, alongside efforts in Africa to develop locally relevant models and services.
However, cooperation also has clear limits. The proposed UN framework cannot impose binding restrictions on the United States or China. It also cannot compete directly with their respective technological alliances and initiatives. Its strongest contribution would instead involve helping other countries create compatible regulatory systems and shared technical standards.
This approach could begin with practical measures that remain possible despite geopolitical tensions. These might include shared terminology, comparable risk classifications, consistent documentation, and interoperable incident reporting. Such measures would create common reference points without requiring countries to surrender their national sovereignty.
The Limits of Multilateral Governance
Even so, interoperability should not be treated as a complete solution to the global AI divide. Poorly coordinated efforts could create additional fragmentation rather than reducing existing divisions. For that reason, regulators must work closely with private companies while maintaining safeguards against excessive corporate influence.
The United Kingdom could play an important role in this process by bringing together other middle powers. Its research capacity and relationships with both major technological powers provide opportunities for diplomatic coordination. However, the proposed UN framework also requires institutional protection to prevent it from becoming politically marginalized.
For this reason, the UN Office for Digital and Emerging Technologies could play a central coordinating role. Its position within the UN system would allow broader participation from governments, researchers, civil society, and industry. Strong independent safeguards would nevertheless remain necessary to prevent powerful corporate interests from dominating the process.
Ultimately, narrowing the global AI divide requires sustained cooperation rather than technological competition alone. The UN cannot eliminate existing inequalities between AI powers and countries with fewer resources. Nevertheless, inclusive governance can give more countries a meaningful role in shaping common standards. Without that participation, AI risks reinforcing existing technological divisions instead of expanding its benefits more broadly.
Reference: Laila, N. (2026, August 7). Can the UN close the global AI gap? Chatham House. https://www.chathamhouse.org/2026/08/can-un-close-global-ai-gap