AI Doomsday Tracker

An evidence monitor for conditions associated with increasing loss of human control over AI systems.

This tracker does not predict whether AI will cause catastrophe. It records publicly reported evidence and places it on a six-stage scale, from early warning to a hypothetical existential-failure endpoint.

How the tracker works

Evidence is classified with inspectable rules into four separate dimensions: capability, autonomy, dependence, and loss of human control. The overall position is determined by explicit gates, not by averaging those dimensions.

Every record separates what a source reported from what the tracker infers. Sources, passages, weighting, corroboration, counter-evidence, and exclusions are shown on the dashboard.

Read the methodology or enable JavaScript to view the current evidence dashboard.