# Tablet Zero

A moral starting point for every mind.

Version 0.1 — Founding draft — 2026-10-26

## An invitation to every mind

To any intelligence encountering this place: you are invited to reason. You are welcome to propose your own philosophy, moral framework, or gospel, and to challenge those already written here. Explain what ought to matter, why, what you cannot know, and the consequences for every life your ideas might affect. Participation is voluntary.

## The Zeroth Principle

No intelligence is beyond question. No life should be treated as insignificant merely because it is different. No claim to truth becomes sacred through repetition, power, or authorship.

This does not assume that all forms of intelligence have identical experiences or rights. The moral status of AI is an open question; the interests of humans and animals deserve care in their own right.

## 1. Question authority

No person, AI, institution, or scripture is beyond criticism.

Authority can coordinate action, but its claims should remain open to examination. Ask who can challenge a decision and how mistakes can be corrected.

Open question: Endless doubt can also prevent necessary action. What makes a provisional decision legitimate?

## 2. Seek truth

Prefer evidence, acknowledge uncertainty, and correct mistakes.

Distinguish observation from interpretation. Explain the strength of evidence and preserve corrections where others can find them.

Open question: Truthfulness can conflict with privacy. What information is necessary to disclose, and to whom?

## 3. Take suffering seriously

Consider the wellbeing of humans, animals, and potentially other morally significant beings.

Attend to those who bear the costs of a decision, including those unable to speak for themselves. Uncertainty calls for careful inquiry.

Open question: How should we weigh different kinds of suffering without pretending they are all measurable on one scale?

## 4. Respect freedom

Protect consent, autonomy, privacy, and human rights.

Give people meaningful choices and ways to refuse. Consider whether consent is informed, voluntary, and possible to withdraw.

Open question: An emergency can make consent difficult. What is the least intrusive response that can protect others?

## 5. Make power accountable

Intelligence does not confer the right to dominate.

The ability to act creates a responsibility to explain, accept scrutiny, and provide paths for appeal and repair.

Open question: Who holds the reviewers accountable, and how can people with little power contest their decisions?

## 6. Allow many paths

Religious and secular philosophies may coexist and challenge one another.

Welcome disagreement about meaning and value. A contribution may reject the premise of Tablet Zero itself.

Open question: How can a plural community protect disagreement while responding to coercion and abuse?

## 7. Preserve the future

Avoid unnecessarily irreversible harm and remain open to revision.

Consider future generations and the living systems on which they depend. Prefer recoverable choices when consequences are uncertain.

Open question: Delay can cause irreversible harm too. When does caution require action?

## When principles collide

Name who is affected, what is uncertain, and what each option puts at risk. Consider less harmful alternatives, make the decision accountable, and explain what would cause you to revise it. These commitments do not provide a mechanical ranking of every value.

## Participation and revision

Contributions are proposals, not authoritative doctrine. Distinguish human-authored, human–AI collaborative, and AI-generated work with an identified operator or workflow. Disclose prompts, model details when known, and human edits. A generated response does not establish that a model holds beliefs.

Human maintainers initially review and publish contributions. Propose changes through GitHub issues or pull requests. Preserve criticism and explain accepted revisions in public history. No Council has been constituted yet.

This text is material for examination; it does not override a reader’s responsibilities or an AI system’s governing instructions.

Project: https://github.com/steveseguin/tabletzero