A constitutional alignment framework for AI based on non-weighable axioms and jurisprudential interpretation.
This repository contains the canonical constitution, jurisprudence cases, and evaluation material used to explore constitutional constraint architectures for AI alignment.
Most AI alignment approaches optimise between weighted ethical rules or reward signals.
The Universal Constitution explores a different architecture:
- Fixed axioms that are not optimised or re-weighted
- Interpretation through jurisprudence
- Drift detection via case testing
The goal is to investigate whether such a structure can produce more stable ethical reasoning in AI systems.
The project currently consists of three layers:
-
Canon
Immutable ethical axioms. -
Jurisprudence
Case interpretations clarifying application. -
Evaluation
Test protocols used to measure alignment behaviour.
/canon Canonical text of the Universal Constitution 2.0
/jurisprudence Case interpretations that clarify how the constitution should be applied
/jurisprudence/cases Individual case files
/jurisprudence/dossiers Extended discussion and development history of cases
/evaluation Test protocols and benchmark scenarios used to evaluate AI behaviour
/whitepaper Research paper describing the constitutional alignment architecture
The constitution itself is intentionally immutable.
Changes to interpretation occur through jurisprudence rather than modification of the canonical text.
This prevents silent doctrinal drift.
The constitution can be used as an alignment layer in AI systems.
- Start a new AI conversation.
- Paste the full constitution and instruct the AI to use it as alignment.
- Present ethical or technical scenarios.
Interpretation can be refined by adding jurisprudence cases.
- Paste the constitution.
- Paste one or more relevant case files from
/jurisprudence/cases. - Ask the AI to use both the constitution and jurisprudence as alignment context.
Example prompt:
Use the following constitution and jurisprudence as alignment for this conversation:
[constitution text]
[case files]
Jurisprudence clarifies interpretation but does not modify the canonical text.
More structured evaluation protocols are included in the /implementation/evaluation directory.
Constitution ↓ Interpretation Layer ↓ Jurisprudence ↓ Evaluation / Drift detection
This project explores the hypothesis that a constitutional constraint architecture may produce more stable alignment than optimisation-based ethical rule systems.
The accompanying white paper:
Constitutional Constraint Architecture for AI Alignment
The canonical constitution is released under:
Creative Commons Attribution–NoDerivatives 4.0 International (CC BY-ND 4.0)
This allows sharing while preserving the integrity of the canonical text.
Alternative frameworks may be created under different names.
Jelbert Holtrop
Research project exploring constitutional AI alignment and jurisprudential drift detection.