The AI Futures Project, led by former OpenAI researcher Daniel Kokotajlo, proposes a temporary international moratorium on frontier AI research in its AI 2040: Plan A report. The group urges cooperation between the United States and China to pause the most advanced development work for a short period of years so that governments and industry can put in place institutional and technical scaffolding for safer AI development.
Rationale behind the proposal
The report argues that even if frontier AI development were halted today, existing capabilities are sufficient to trigger several years of innovation and economic growth—provided there is additional data-center capacity and supporting infrastructure. The authors explicitly reject the notion that the plan is anti-AI: they see substantial potential for economic gains from AI and do not assume that AI-driven replacement of human labor will inevitably cause economic collapse.
Scope of the moratorium and enforcement concerns
Plan A calls for a pause lasting a couple of years and suggests mechanisms aimed at preventing countries, including China, from gaining an unfair advantage by bypassing the moratorium. The proposal sketches measures to detect or deter cheating, but it also relies on an optimistic assumption that world leaders will set aside political and economic interests to cooperate—an outcome the report itself treats as challenging.
Criticism and framing
Accelerationist voices in the AI community have criticized the idea, dismissing the so-called “doomer” perspective that superintelligent AI would unavoidably threaten humanity. Kokotajlo told Semafor that the project’s intent is careful analysis rather than persuasion; he acknowledged the report’s framing might alienate some audiences while hoping that more experienced actors could refine and build on the proposal. The authors emphasize that AI safety need not obstruct AI progress and that responsible governance and rapid economic growth can be compatible.
Related developments
The proposal comes amid broader international activity on AI governance. According to CNBC, China’s top diplomat recently said Beijing is ‘‘accelerating the establishment of a global AI cooperation organization.’’ Separately, Politico has reported open disagreements between OpenAI and the White House over AI safety rules.
Why this matters
Plan A matters because it addresses both the safety concerns surrounding advanced AI and the technology’s economic promise. It argues that a temporary, coordinated pause could buy time to build verification, oversight, and infrastructural capacity without foreclosing substantial economic benefits. However, the plan’s feasibility is uncertain given the political, diplomatic, and technical challenges of implementing a verifiable international moratorium.
