Papers & Artifacts

Latest versions and archival releases.

• The Don’t-Be-a-Dick Ethical Framework (v2.0)
• The Don’t-Be-a-Dick Ethical Framework (v1.0)

Abstract (v2.0)

We propose a minimal decency metric E(A) combining five bounded inputs—harm, autonomy/consent, intent, proportionality, and transparency—with interpretable thresholds for action. We outline a preregistered program to evaluate cross-cultural agreement and domain-specific weighting, and discuss failure modes and falsification criteria.

Method sketch

1. Design multilingual vignettes spanning everyday, civic, and AI-mediated actions.
2. Collect ratings for H, C, I, P, T and overall judgment; hold-out sets for validation.
3. Fit weight vectors per domain; test stability across cultures; publish preregistered thresholds.

Data releases will appear here as rolling CSVs and figures.