If you only read one thing: RippleLogic is a filter-first decision process. It eliminates unsafe options before scoring “best.”
A safety-first “governance seatbelt” for high-stakes decisions:
reject rights violations → bound catastrophic risk → protect system integrity → then optimize welfare.
RippleLogic is a safety-first decision framework for humans and AI systems making consequential choices.
It works in a strict order:
- Reject any option that violates basic rights.
- Rule out options with unacceptable catastrophic tail risk.
- Protect the structural integrity of the systems that must keep functioning.
- Only then rank remaining options by their net welfare impact.
Stakeholders are modeled as seven nested unions (Self through Biosphere).
Impacts are recorded in a transparent 49-cell matrix (7 unions × 7 welfare dimensions).
RippleLogic prevents “good outcomes” from being used to justify unacceptable means.
It blocks rights-violating or catastrophe-prone options first, and only then compares what remains.
RippleLogic is built for verification, not belief. You do not have to trust it. You can inspect how it reached a conclusion.
- Evaluation scope: Claims are limited to Tier 1–3 evaluation.
- Auditability: Outputs are designed to be reconstructible by independent reviewers.
- Falsifiability: Explicit criteria define what evidence would show the framework fails.
No Tier-4 determinism or predictive governance guarantees are claimed.
- Release page: https://github.com/MathGov/ripple-logic/releases/tag/v8.1
- Canonical release folder:
releases/v8.1_2026-02-14/
That release contains:
- Core specification (RippleLogic v8.1)
- Sentience Gradient Protocol (SGP v4.2.3)
- Agent System specification
- Worked exemplar (Ripple Aligners Tier-2)
- ProofPack skeleton / verification artifacts
- SHA256 integrity file (checksums)
- Download the release assets (or the full bundle, if provided).
- Locate the SHA256 checksums file in the release artifacts.
- Compute SHA256 locally and confirm it matches the published checksums.
(If you publish a one-command verification script later, link it here.)
Project home and public docs: https://ripplelogic.org
See CONTRIBUTING.md.
Please use Issues for:
- clarity gaps / confusion points (especially “what did you mean by X?”)
- test vectors and counterexamples
- falsification attempts
- pilot deployment notes
See SECURITY.md for how to report vulnerabilities or safety issues.
Apache-2.0 (see LICENSE).
In one sentence:
RippleLogic prevents rights violations and catastrophic decisions before optimizing outcomes.