Twenty governments and the EU adopted a formal call for frontier AI controls in October 2026, while the United States declined to join a multilateral governance framework. In the same window, senior ministers called for the first time for formal negotiations toward an autonomous weapons treaty, a notable escalation from prior years' more general statements of concern.
The split between the signing bloc and the US isn't simply a disagreement about AI risk; it reflects a deeper divide over whether frontier AI capability is primarily a national competitive asset to be protected or a shared global risk to be jointly managed, and that framing difference has shaped nearly every multilateral AI governance effort since 2023.