AI Futures Project is the research organization behind *AI 2040: Plan A*, written by Thomas Larsen, Romeo Dean, Brendan Halstead, Eli Lifland, Ryan Greenblatt, and Daniel Kokotajlo. The project uses detailed scenarios to make assumptions and causal chains in AI policy proposals easier to inspect than abstract recommendations alone.source: ai-2040-plan-a-2026.md
The authors present Plan A as a recommendation and stress test, not their most likely forecast. It follows their earlier *AI 2027* scenario but asks what a deliberately better transition could look like: a verified international slowdown, broad research transparency, diffusion of frontier capability, and a staged path toward superintelligence.source: ai-2040-plan-a-2026.md
The organization’s method is scenario scrutiny: write a concrete sequence in which a proposal succeeds, then examine implementation difficulties, incentives, side effects, and failure modes. This is intended to expose load-bearing assumptions that policy slogans leave implicit.
Related pages: ai-2040-plan-a, ai-compute-infrastructure, mutually-assured-compute-destruction, ai-labor-market.