Iceberg Layer
Mental Models
Structures
Patterns
Events
Boundary / Exogenous
Feedback Loops
Closed causal chains that reinforce (R) or balance (B) themselves. The core dynamic structures of the system.
Forcing Chains
Open causal chains missing one closure edge. The gap reveals where the system has no self-correcting mechanism.
Harm Mechanism
ELL disparate-impact subgraph — 5 vulnerability channels converging on ELL students through AI detection, locked in a reinforcing loop.
IXDS 709 · Systems Analysis · SCAD 2026
The Fix That
Makes It Worse
A causal loop analysis of how AI is reshaping trust, skill, and assessment in art + design education — and why the instinctive institutional response amplifies every harm.
88 variables·188 edges·13 feedback loops·3 control clusters
Click to enter↓