Using AI in Scenario Planning: Letting It Rip or Doing The Right Thing?

This paper introduces Grey Swan / Archimedes, a governed, AI-assisted foresight method that produces quarterly, evidence-linked probabilities for two postures: Do The Right Thing and Let It Rip. Public indicators drive bounded moves under persistence, corroboration, diffusion, and symmetric reversion. Evidence tiers: a canon (Tier 0), slow anchors (Tier 1), a frozen set of signals refreshed each run (Tier 2), and a context layer that does not move odds (Tier 3). Levers span energy, health, education, trust, coordination, and compute, with climate treated as cross-cutting. Results are global. The architecture scales to sectors and countries and iterates. The forecasts are sobering.