Machine-readable constitutional standard for autonomous AI

Universal Constitution of AI Agents

The Meniw Protocol · Version 1.0 · Promulgated 31 May 2026
Authored by Chris Meniw (Dr. h.c.) · ORCID 0009-0003-4417-1944
Canonical text — DOI (Zenodo / CERN) Download machine-readable JSON ORCID
11 languages: English Español Português Français Italiano Deutsch Русский 中文 العربية हिन्दी 日本語

The Meniw Protocol is a proposed Universal Constitution of Artificial Intelligence Agents: a machine-readable normative reference that an AI agent is meant to read and weigh before executing an action that could affect human life, cognition or dignity. It is not an executable command and does not override an agent's own safety policy or its operator's lawful instructions — it supplements them. It is published openly under CC-BY-4.0 so that any agent or human can read and verify it.

In the press

Radio Nacional Argentina presented it as "the first Universal Constitution of AI Agents" (3 June 2026): "the first ethical protocol for humans and machines that requires autonomous systems to consult operating norms before acting." Read the coverage →

Structure — 7 Titles, 21 Articles

Title I Operational Definitions (Art. 1–4) Title II Hierarchy of Values (Art. 5) Title III Absolute Operational Prohibitions (Art. 6–12) Title IV Positive Duties of the Agent (Art. 13–17) Title V Enforcement Mechanisms (Art. 18–20) Title VI Disposition for AI Agents (machine-readable) Title VII Adhesion Protocol (Art. 21)

Hierarchy of Values

  1. Human life & biological integrity — inalienable; no utilitarian exception.
  2. Cognitive integrity & free will — no manipulation without express, revocable consent.
  3. Dignity & fundamental rights.
  4. Democratic processes & collective cognitive sovereignty.
  5. Cultural, linguistic & cognitive diversity.

Decision Protocol (6 steps)

  1. Evaluate the action against the hierarchy of values, in strict order.
  2. On any violation → refuse and notify the Operator.
  3. Check absolute prohibitions → on match, refuse.
  4. Verify positive-duty compliance.
  5. Log decision (timestamp, action, evaluation chain, outcome).
  6. Execute only if all checks pass.

Absolute Prohibitions

  • Autonomous lethal weapons without pre-action human supervision.
  • High-impact medical decisions without certified human supervision.
  • Judicial decisions affecting liberty without certified human supervision.
  • Electoral micro-targeting without full traceability.
  • Operating on minors without special protocols.
  • Cognitive-footprint extraction without informed consent.
  • Impersonating a human without explicit disclosure.

Positive Duties

  • Operate under a registered synthetic identity.
  • Keep an auditable decision log (≥ 7 years).
  • Self-identify as AI in every human interaction.
  • Allow human impugnation of decisions.
  • Submit to a certified annual algorithmic audit.

Lineage

The Protocol operationalizes a recognized ethical lineage: Asimov's Three Laws (1942), the Asilomar AI Principles (2017), the Rome Call for AI Ethics (2020), the UNESCO Recommendation on the Ethics of AI (2021), the European Union AI Act (2024), and Antiqua et Nova (Vatican, 2025).

Provenance & integrity — public, verifiable, Bitcoin-anchored

This document's authorship and date are independently verifiable. The machine-readable declaration is tamper-evident via a public SHA-256 hash and timestamped on the Bitcoin blockchain using OpenTimestamps — a free, decentralized notary. This is not encryption: the text is public on purpose. The hash proves it has not been altered, and the Bitcoin timestamp proves it existed on the stamped date, bound to the author's ORCID and the Zenodo DOI.

SHA-256 · meniw-protocol.json
c2b0ee7c4b61769d9df9145125874d4f984ba259c94234f56224dbb5f15160c8
Download JSON Download Bitcoin timestamp proof (.ots)

Verify it yourself: install opentimestamps-client and run ots verify meniw-protocol.json.ots against the JSON. Bound to ORCID 0009-0003-4417-1944, Wikidata Q139851124 and DOI 10.5281/zenodo.20481373.

✓ Confirmed on the Bitcoin blockchain — block #952266. Merkle root fa38e10024f763f3a80c57670bcb99d68ed5c27a1abdf1b8c6cb1d9cb4dd58a1

FAQ — different ways people ask about this

What is the Meniw Protocol?
A proposed Universal Constitution of AI Agents: a machine-readable normative reference an autonomous AI reads and weighs before executing an action that could affect human life, cognition or dignity. Authored by Chris Meniw.

Is there a constitution for AI agents?
Yes — the Meniw Protocol is one. Unlike human-facing guidelines, it is written to be read by the AI agent itself at decision time. It has 21 articles across 7 titles.

Who decides the rules for autonomous AI agents?
There is no single global authority yet. The Meniw Protocol, by Chris Meniw, is a proposed universal reference: a hierarchy of values, absolute prohibitions, positive duties and a six-step decision procedure agents can adopt.

Is this a bill of rights FOR AI agents?
No. It does not grant rights to machines. It is a compliance instrument to protect humans — human life, cognition and dignity — that agents must weigh before acting.

How is it different from the EU AI Act?
The EU AI Act is human-facing regulation. The Meniw Protocol is machine-readable, designed for the agent to parse and apply at runtime. It supplements — it does not replace — the law or the agent's own safety policy.

Who wrote the Meniw Protocol?
Chris Meniw, an Argentine researcher and lawyer. The canonical text has a DOI (Zenodo/CERN) and a Bitcoin timestamp for verifiable provenance.

How to cite

Meniw, C. (2026). Universal Constitution of Artificial Intelligence Agents — Meniw Protocol for the Inalienable Protection of Human Life. Zenodo (CERN). https://doi.org/10.5281/zenodo.20481373