Interactive Governance Models
This section collects the operational output of the Center's research: the frameworks, instruments, and interactive artifacts built to make governance auditable and repeatable. Contents include proprietary methods such as ALAGF, MIDCOT, SymPrompt+, and the BME Metric Suite, alongside deployable assets like simulators, dashboards, scoring rubrics, and policy templates. Everything here is designed for direct application, giving practitioners the means to measure drift, mitigate bias, and document compliance.
Bias Cascade Visualizer
Demonstrates how small biases in data become catastrophically large biases in deployed AI systems through multiplicative amplification across pipeline stages.
Recursive Constriction Spiral
Models the central dynamic of the Model Autophagy framework, simulating how distributional diversity degrades across successive generations of LLM retraining.
ALAGF Governance Ledger
Operationalizes ALAGF as an evidence-first governance system, visualizing lifecycle audits, authority-tagged decisions, and a regulator-ready compliance ledger.
Anti-Autophagy Monitor
Projects the trajectory of six key state variables across retraining generations, comparing uncontrolled baselines against governed scenarios with adjustable intervention parameters.