Human-AI Relations — a working diagram

The Loop Behind the Model

This isn't a diagram of a prediction. It's a diagram of the process that produced one — symbolic instinct, structured into a model, run against real facts by one AI, checked for its epistemic honesty by another, and returned as a dated, public, falsifiable trail. The loop doesn't end in an answer. It ends back at a person deciding what to make of it.

Instinct human Model structure Generation AI engine Verification AI engine Trail record Reader human
Select a stage

How this reads

Five stages plus the reader, arranged as a loop rather than a line. The two human-touched stages — Instinct and Reader — are marked in blue. The three AI/model stages in between are brass, the same palette as the Gears diagram.

Follow the loop clockwise from Instinct. The one thing to notice: it doesn't terminate at Trail. It returns to a person. Nothing in this loop claims the last word.

Click any stage to open its detail →
Human-touched stage Model / AI stage
Reading note — this diagram documents a real, already-run process — it isn't a proposal for how human-AI interaction should work in general, just an honest map of how this specific model got tested. See the Testing Portal for the actual run this diagram is describing.
→ Run the Workflow — a guided tool to run this loop yourself, across three AI engines, on any topic you choose.