Article Series
The International AGI Project Series
Released on 26th January 2026
Citations
Introduction
This is a series of papers and research notes on the idea that AGI should be developed as part of an international collaboration between governments.
Most of this work was written as part of a research avenue that we don’t currently plan to pursue further. It’s more like work-in-progress than Forethought’s usual publications, but we’re sharing it as we think some people may find it useful.
We wanted to (i) assess how desirable an international AGI project is; (ii) assess what the best version of an international AGI project (taking feasibility into account) would look like.
The result is the proposal described in “What an international project to develop AGI should look like.” The core idea is that we can get most of the benefits of an international project by giving non-US countries meaningful influence over only a relatively small number of decisions. By making non-US influence circumscribed in this way, and letting the US call the shots day to day, the proposal becomes both more feasible and less likely to get bogged down in bureaucracy.
The proposal is modelled on Intelsat, an international project that developed the first global satellite communications network, so we call this the “Intelsat for AGI” plan. Intelsat is explained and discussed further in “Intelsat as a Model for International AGI Governance.”
The other research notes supplement this core proposal:
- “AGI and World Government” argues that the possibility of an intelligence explosion makes an international AGI project more desirable.
- “International AI Projects and Differential AI Development” covers an important way in which any such project should be circumscribed, if possible.
- “A Global Convention to Govern the Intelligence Explosion” shares a plan that could make an international AGI project more likely (and which seems independently desirable).
Two other pieces of background research into existing institutions indirectly informed our discussion of this area. These are:
All of the pieces can be read in isolation, or the series can be read in order.
Released on 26th January 2026
Citations
Articles
What would the best version of an international project to develop AGI look like? In this research note, I set out my tentative best guess: “Intelsat for AGI”. This would be a US-led international project modelled on Intelsat (an international project that set up the first global communications satellite network), with broad benefit sharing for non-members. The primary case is that, within the domain of international AGI projects, this looks unusually feasible, and yet it would significantly reduce catastrophic risk compared to a US-only project.
If there is an international project to build artificial general intelligence (“AGI”), how should it be designed? Existing scholarship has looked to historical models for inspiration, often suggesting the Manhattan Project or CERN as the closest analogues. But AGI is a fundamentally general-purpose technology, and is likely to be used primarily for commercial purposes rather than military or scientific ones.
This report presents an under-discussed alternative: Intelsat, an international organization founded to establish and own the global satellite communications system. We show that Intelsat is proof of concept that a multilateral project to build a commercially and strategically important technology is possible and can achieve intended objectives—providing major benefits to both the US and its allies compared to the US acting alone. We conclude that ‘Intelsat for AGI’ is a valuable complement to existing models of AGI governance.
If there’s a large enough intelligence explosion, the first project to build AGI could organically become a de facto world government. In this note, we consider what implications this possibility has for AGI governance. We argue that this scenario makes it more desirable that AGI be developed by a multilateral coalition of democratic governments, under explicitly interim governance arrangements, and that non-participating countries receive major benefits and credible reassurances around their sovereignty.
Proposals for an international AI project to manage risks from advanced AI generally require all frontier AI development to happen within that project, and with limitations. But some AI capabilities actively help with addressing risk. I argue that international projects should aim to limit only the most dangerous AI capabilities (in particular, AI R&D capabilities), while promoting helpful capabilities like forecasting and ethical deliberation. If this is technically feasible (which I’m uncertain about), it could increase our capacity to handle risk, reduce incentives to race, and help get industry on board with an international project.
There currently isn't a plan for how society should navigate an intelligence explosion. This research note proposes an international convention triggered when AI crosses defined capability thresholds. At that point, the US would pause frontier AI development for one month and convene other nations to draft treaties to govern the many challenges an intelligence explosion would throw up. While potentially feasible if agreements can be made quickly enough, it’s unclear if enforcement and technical details would work in practice.
This rough research note gives an overview of some international organisations and their voting structures, as background for thinking about the international governance of AGI.
The transition to advanced AI systems may eventually lead to some kind of international agreement to govern AI. An important historical case study for an agreement of this kind is the founding of the United Nations. This research note gives an overview of the creation of the UN charter, before drawing some tentative conclusions for international AGI governance.