This project is part of the broader ANGEL Project initiative, which aims to bridge the gap between complex digital systems and human well-being. The Cultural Dissonance Detector (CDD) is a machine learning–based simulation tool that identifies patterns of cultural contradiction and stress in AI outputs or institutional discourse.
The detector simulates how a model or system responds when it encounters implicit cultural contradictions—statements or practices that appear normalized but conflict with empirical reality, scientific sustainability, or collective human thriving.
- Dissonance Scoring: Quantifies contradictions between profit-seeking, sustainability, and dignity-based values.
- Validation Loop: Tests AI and institutional outputs for embedded stress patterns, inconsistency, or maladaptive assumptions.
- Ethical Signal Detection: Flags narrative distortions that perpetuate economic or ecological harm under the guise of stability.
- AI alignment research and interpretability
- Institutional reform simulations
- Evaluating educational, corporate, or governmental messaging
- Tools for psychologists, sociologists, and cognitive scientists
notebooks/: Jupyter notebooks for simulation and scoring logicdata/: Placeholder for test content (e.g. policy briefs, corporate statements)models/: Optional future expansion for fine-tuned modelsresults/: Output from simulations
- Finalize and run the prototype notebook
- Benchmark known contradictory narratives (e.g. greenwashing, military humanitarianism)
- Validate the scoring framework for academic and real-world application
Maintainer: Robin Macomber
Support AI: Numin
Part of: ANGEL Project