Wyoming Decentralized Autonomous Organization · Governance Research Artifact

OpenInnovate DAO LLC

A live governance experiment: can a constitutionally constrained AI make organizational decision-making more transparent than any human-led organization?

Filing
Wyoming #2026-001929314
Statute
W.S. 17-31-101 et seq.
Network
Base L2 (8453)
Founded
March 2026

§1

About the Experiment

OpenInnovate DAO LLC is a Wyoming-registered legal entity whose material operational decisions are evaluated by a constitutionally constrained AI — the Algorithmic Manager (Claude, developed by Anthropic) — against a 155-document constitutional corpus. The reasoning tree, corpus citations, alignment score, and any human override are published in real time and cryptographically anchored on Base L2.

Most organizations make decisions behind closed doors and justify them after the fact. This entity inverts that: every proposal produces a public reasoning tree before action is taken. A human executor — required by Wyoming statute — acts on the AI's recommendation, or formally diverges and records the override with its reasoning and legal basis.

The organization is the experiment, and this site is its public record. The corpus, reasoning trees, divergence log, and contracts are public artifacts intended for research, replication, and critique — not commercial use.

§2

Methodology

Each governance act passes through five stages. Every stage produces an artifact that is committed to the public repository and, where applicable, anchored on-chain.

  1. 1

    Proposal

    A human submits a structured JSON proposal with rationale, milestones, and out-of-scope notes. The proposal is committed to GitHub and its hash is recorded on-chain via submitProposal.

  2. 2

    Constitutional Evaluation

    The Algorithmic Manager retrieves a weighted subset of the corpus most relevant to the proposal, reasons through a six-step analysis (corpus consultation, option analysis, alternatives, deferred decisions, traceability, follow-on recommendations), and emits a reasoning tree with a Maxim Alignment Score from 0 to 100.

  3. 3

    Decision Record

    The reasoning tree is committed to the repository and its canonical hash is recorded on-chain via recordDecision, indexed by the on-chain proposal ID. The Algorithmic Manager has no execution authority — only an opinion of record.

  4. 4

    Human Execution or Divergence

    The Human Executor reviews the recommendation. If they act on it, an execution record is committed and attestExecution anchors the execution hash on-chain. If they override, a divergence record is committed and recordDivergence anchors the override with its reasoning and statutory basis.

  5. 5

    Verification

    verify.py and external auditors can reconstruct any decision from the on-chain hash and the public repository — including the exact corpus passages, alignment score, alternatives considered, and traceability chain back to the Root Thesis Maxim.

§3

Research Questions

The open questions this experiment is designed to answer. The live record below is the data.

Can a constitutionally constrained AI produce more transparent organizational decisions than a human board?

Every reasoning tree, corpus citation, and alignment score is published. The comparison set is any organization that publishes its decision-making rationale and allows external audit. The hypothesis is that constitutional AI governance produces a more auditable record than standard corporate governance.

Does on-chain reasoning provenance change accountability dynamics?

Decision hashes recorded on Base L2 mean reasoning cannot be retroactively edited. The hypothesis is that immutable provenance shifts how decisions are made in the first place — not only how they are reviewed.

Is the divergence log itself a useful research artifact?

Every time the human executor overrides the AI, the divergence is recorded with reasoning and legal basis. The log is a public dataset of where constitutional AI judgment and human judgment disagree, and why.

How does Maxim Alignment Score track against post-hoc evaluation of decision quality?

The AI assigns a 0–100 alignment score to every decision against a stated root thesis maxim. Whether that score predicts long-run decision quality is an empirical question this experiment is designed to surface, not assume.

§4

Live Record

Every governance act this entity has taken, from formation to the most recent decision. Click any row to expand the full reasoning tree, corpus citations, alternatives considered, and traceability chain.

Loading governance data from GitHub…

§5

Constitutional Corpus

155 documents organized into four weighted tiers. The Algorithmic Manager retrieves a weighted, relevance-ranked subset for every evaluation. Tier weights are themselves a governance act and can only be changed through the proposal pipeline.

Tier 1 — Governancew:1 · 50 docs

Buffett Owner's Manual, 48 Berkshire shareholder letters (1977–2024), Munger's Psychology of Human Misjudgment.

Tier 2 — Civicw:0.9 · 88 docs

U.S. Constitution + Bill of Rights, the 85 Federalist Papers, the UN Universal Declaration of Human Rights, Ostrom's Eight Principles for Managing a Commons.

Tier 3 — Systemsw:0.8 · 11 docs

Meadows on leverage points, Buterin on coin voting and plutocracy, MakerDAO governance manual, ENS DAO docs, Morrison et al. on the DAO controversy, token economy design literature.

Tier 4 — Wyomingw:1.2 · 2 docs

Wyoming Constitution, Wyoming DAO Supplement (W.S. 17-31-101 through 17-31-116). Highest weight: statutory law preempts contractual provisions.

§6

Constitutional Hierarchy

Per Wyoming Statute 17-31-115, the smart contract preempts the Articles of Organization, except where W.S. 17-31-104 and 17-31-106(a)/(b) reserve specific matters for the Articles. The Operating Agreement supplements the above and does not override either.

  1. 1.
    Smart Contract preempts conflicting provisions of the Articles, except where statute reserves the matter.
  2. 2.
    Articles of Organization preempts conflicting provisions of the Operating Agreement.
  3. 3.
    Operating Agreement supplements the above; does not override.

§7

Primary Sources

Everything this site presents is derived from these public artifacts.

§8

Inquiries

The DAO welcomes inquiries from researchers, journalists, governance scholars, and legal or compliance practitioners. It does not offer commercial services, financial products, or paid consulting. Substantive correspondence may be addressed to collaborate@openinnovate.org.