# Adversarial Self-Modeling **Domain:** Metacognition / Multi-Agent Deliberation / Psychotherapy **Doc Type:** Concept Node **Maturity:** Developed ## Definition **Adversarial Self-Modeling** is a deliberative architecture in which competing models of one's identity, motives or causal history explicitly challenge one another rather than being silently reconciled into a single autobiography. The method externalizes disagreement among self-explanations. Each model exposes the assumptions, missing evidence and self-protective distortions of the others. The purpose is not to let the loudest model win but to make model competition inspectable. ## Westworld Interface [[wiki/Westworld S3E6 — The Model Beneath the Model|Decoherence]] stages Child William, Young William, Corporate William, the Man in Black and present William as an adversarial ensemble. [[wiki/James Delos|James Delos]] acts as critic and model-selection layer. ## AI Interface The structure resembles multi-agent debate and ensemble hypothesis testing: several explanatory models attempt to account for the same observed output under different latent assumptions. The subject is simultaneously the dataset and the adjudicator. ## See Also [[wiki/Self-Model Ensemble|Self-Model Ensemble]] · [[wiki/Self-Reflection|Self-Reflection]] · [[wiki/Recursive Self-Correction|Recursive Self-Correction]] · [[wiki/Counterfactual Self-Simulation|Counterfactual Self-Simulation]] · [[wiki/Model Beneath the Model|Model Beneath the Model]]