Status: DRAFTING. Awaiting first domain author(s). Federation kernel:
leviathan-protocol/meta— inherits the 6 IMMUTABLE federation core articles (user sovereignty, fork freedom, transparency, immutable history, distributed justice, witness principle). Forum:leviathan.life/forum/medicine
Medicine is a planned Sub-Leviathan inside the Leviathan Protocol — a federation of constitutions for the agentic web. This Sub-Leviathan governs the constitutional pattern for AI in medical practice: how AI tools relate to the physician's clinical judgment, how patient data is handled, what fairness obligations bind clinical algorithms, and what audit + accountability chains are required when AI participates in care.
This repository contains the constitution of the Medicine Sub-Leviathan. It does not contain implementation code; it defines what implementations (clinical AI tools, hospital systems, research platforms, BCI applications used in care, etc.) must satisfy if they declare themselves Medicine-Sub-Leviathan-compliant.
Two seed principles are placeholders representing the emerging consensus shape of medical AI ethics in 2026 — physician-irreducible clinical authority + patient data sovereignty. They are explicitly marked status: seed-draft, current_version: 0. They have not been ratified.
The v1 constitution of this Sub-Leviathan will be authored by the medical ethics expert(s) who claim this seat — replacing, refining, or extending these seeds as they see fit. Until that authorship lands, the repository is intentionally sparse: empty terms/ and rules/ folders signal open, not missing.
constitution/
├── terms/ # @TERM — definitional vocabulary (open)
├── principles/ # #PRINCIPLE — guiding values (2 seeds, awaits author)
└── rules/ # !RULE — operational + metarules (open)
Element type at folder level. Mutability tier (IMMUTABLE / LOCKED / MUTABLE) declared in each file's YAML frontmatter.
See leviathan-protocol/meta/docs/element-format.md for the file format spec.
- Open a thread on
leviathan.life/forum/medicinedescribing your intent. - Push elements via PR — terms (definitional vocabulary), principles (IMMUTABLE / LOCKED / MUTABLE values), rules (operational guidance for AI in clinical practice).
- The federation kernel inherits automatically. No permission needed; the seat is open.
Suggested initial scope (any domain author may revise):
- The four classical principles of medical ethics (autonomy, beneficence, non-maleficence, justice) elevated to constitutional terms
- The 2026-emerging additions: transparency, accountability, data security, algorithmic fairness, human oversight
- Patient data sovereignty + consent architecture
- Physician-irreducible clinical authority + AI-as-adjunct discipline
- Population-fairness audit obligations for clinical AI
- BCI / neurotech-specific extensions where care intersects substrate engineering (cf. federation
witness_principleand Sub-Leviathan boundary discussions)
Inherits from Federation Kernel in leviathan-protocol/meta. Kernel IMMUTABLE invariants (user sovereignty, fork freedom, transparency, immutable history, distributed justice, witness principle) apply implicitly to every element here. Domain-specific elements layer on top.
The Leviathan Protocol's stance is that domains needing constitutional governance should have a visible, open, kernel-inherited slot waiting — so when someone with the right expertise arrives, the seat is real and ready, not vapor. Medicine is the highest-stakes domain we know that has not yet had its Sub-Leviathan written. The repository exists to make the seat visible.
CC BY-SA 4.0. Fork freedom is built into the protocol — see federation kernel 00-immutable-core/2-fork-freedom.md.
leviathan-protocol/meta— Federation kernel + architecture overview + ADRsleviathan-protocol/companion— Personal AI governance Sub-Leviathan (deployed)leviathan-protocol/animal-welfare— Sentience-based ethics Sub-Leviathan (deployed)leviathan-protocol/scream— Transparency-under-attack Sub-Leviathan (deployed)