Oxford Martin School AI Governance Initiative

To reclaim the spirit of the 2024 HLAB‑AI mandate and address the merits of establishing a dedicated Agency for the governance of AI under the auspices of the UN.. Despite the difficulties of the above given the current geo-political landscape I would urge it should at least be on the agenda of the Geneva Global Dialogue meeting.

Interoperability of governance approaches;

Three structural tensions continue to impede progress: geopolitical fragmentation, institutional inertia, and reliance on voluntary mechanisms. Strategic competition between the United States and China further obstructs convergence. Meaningful governance, as envisioned by the HLAB AI, cannot emerge from such a landscape fractured by the strategic gaps that have been outlined here. In this environment strong commercial actors wield disproportionate influence. What is missing is not vision or launch events, but durable institutional design. The time for cautious optimism has passed; what is required now is strategic diplomacy, sustained engagement, and the resolve to move beyond rhetoric into enforceable practice.

A recommendation is to establish a standing mechanism for rapid response to AI related crises, While existing initiatives focus on dialogue, data and standards, none yet provide the capacity to act decisively and with operational depth when unforeseen events occur, whether technical failures, malicious misuse, or geopolitical escalations involving advanced AI systems. A Global AI Risk and Crisis Response Mechanism would fill this gap by offering a structured channel for emergency coordination. The mechanism would draw inspiration from precedents such as the International Health Regulations and nuclear safeguards regimes, combining early warning systems, inspection protocols, and rapid coordination procedures.

The topic of an International AI Agency is not about countries, regions or sectors. It is about created a home where the many AI related governance topics (both on AI safeguards and AI for Good) can start to be gradually and systematically addressed. A body that will orchestrate the myriad of current and emerging governance efforts worldwide.

See previous response regarding an IAIA

The IAEA. Big Science collaborations including CERN and ITER

The rivalry between ODET, ITU, UNESCO and other UN bodies needs to be aside

Bog Tech Academics such as myself who have researched an published on how epistemic communities of experts 'find a way' to form and maintain global cooperation

The current webinars and call for ideas seems the other approach possible - until the co-chairs start to settle on key advisors they would wish to guide implementation rather than just rhetoric.

https://doi.org/10.1093/ia/iiaf105