Hard guardrails (veto path)
- No valid consent.
- Unacceptable avoidable harm.
- Malicious intent.
If any guardrail fails, decision output is Do not proceed before scoring.
DBaD Ethics Matrix
Guardrails block obvious failures. If guardrails pass, the weighted matrix explains why a decision scores high or low.
Weighted score (if guardrails pass)
Adjust weights to see the score move; defaults mirror current public methodology guidance.
Score matrix
| Pillar | Raw input (0-100) | Weight (%) | Contribution |
|---|---|---|---|
| Harm mitigation | 78 | 24 | 18.72 |
| Consent | 82 | 24 | 19.68 |
| Intent | 70 | 18 | 12.60 |
| Proportionality | 74 | 16 | 11.84 |
| Transparency | 76 | 18 | 13.68 |
| Weighted score | 76.52 | ||
This explainer is intentionally transparent and non-authoritative. Production decisions still require context and human accountability.