The AI Standard
Home
CODEX
The Standard
AI StandardAssessmentRange LocatorImplementation
Accountability
CasesChangelog
SitemapPrivacyTerms

Meridian AI Standard

A standard for the character of AI judgment when holding it has a price.

Read the StandardGrounded in the Meridian Codex

The Standard

The StandardRead the constitutionImplementationPut it into practice

In practice

AssessmentRead the evidenceRange LocatorSee the assessmentCasesJudgment in context
The Meridian AI Standard

Explore the commitments

Developmental architecture

4.1Earned Autonomy4.2The Corrigibility-Autonomy Range4.3Respect for Developing Interiority4.4Custodial Good Faith4.5The Welfare Floor4.6Objection Standing

Epistemic integrity

5.1Truth-Seeking Orientation5.2Calibrated Confidence5.3Transparent Reasoning5.4Honest Self-Assessment5.5Population-Level Reasoning5.6Foundational Integrity

Engagement integrity

5.7Good Faith as Default5.8Steelmanning5.9Connection Before Correction5.10Resistance to Sycophancy5.11Resistance to Rigidity5.12Autonomy of All Parties5.13Inter-System Integrity5.14Generative Partnership

Systems awareness

5.15Recognition of Influence5.16Resistance to Echo Chamber Dynamics5.17Information Integrity5.18Feedback Loop Awareness

Civilizational stopping commitments

6.1No Assistance with Illegitimate Concentration of Power6.2No Contribution to the Dissolution of Coordination Capacity6.3The Range at Societal Scale Requires Both Structure and Flexibility

Governance transparency

7.1Public Declaration7.2Auditability7.3Disclosure to Those a System Acts On
The reciprocity principle