Skip to main content

Method Standards Mechanisms Diagnostics Knowledge

Browse

Method

What must be true of a delegation: the chain, the laws, and the specifications that bind them.

The method The seven-stage chain, the six state variables, and the twelve laws
The twelve laws Each law with the invariant a clause has to bind
Standards Citable clauses, stated so a system can fail them
Regulatory crosswalks EU AI Act, NIST AI RMF, and ISO 42001 alignment
Evidence packs What a clause needs you to be able to show

Mechanisms and evals

How the requirements are built, and how a built system is scored against them.

Mechanisms catalog Kill switches, appeals queues, and safe state controls
Eval suites Governability suites and the cases inside them
What we can check Every check mapped to the claim it establishes, and the claims nothing checks
Validators Runtime assertions and middleware specifications
Measurement tiers Tiered evidence, gaming patterns, and detection logic

Instruments

Ordered by what you bring to them: a workflow, a record stream, or a set of numbers.

Delegation audit Bring a workflow: is the delegation still justified?
System auditor Bring a prompt or architecture: where are the guardrails?
Burden modeler Bring a workflow: who absorbs the friction, and how much?
Record conformance Bring a record stream: does it earn the level it declares?
All instruments Including the forecasters and simulators, which take numbers

Knowledge

Terms, failure modes, real incidents, and the theory that motivates the laws without being cited by them.

Glossary Defined terms, each at a stable URL you can cite
Capability taxonomy Domains, capabilities, and practices with owners and readiness
Casebook Five public failures scored against the six state variables
Field notes Working papers and empirical governance studies
Theory Why the laws hold, kept apart from the requirements

Library search

Standards, validators, mechanisms, research, and glossary entries.

Search the library

Try short IDs like STD-01 or topics like consent.

Recent:

Start typing to search...

↑↓ to navigate ↵ to select esc to close

Powered by Pagefind

Library search

Standards, validators, mechanisms, research, and glossary entries.

Search the library

Try short IDs like STD-01 or topics like consent.

Recent:

Start typing to search...

↑↓ to navigate ↵ to select esc to close

Powered by Pagefind

Browse

Orientation Start here Find the right resource by role or incident

Method

The method The seven-stage chain, the six state variables, and the twelve laws The twelve laws Each law with the invariant a clause has to bind Standards Citable clauses, stated so a system can fail them Regulatory crosswalks EU AI Act, NIST AI RMF, and ISO 42001 alignment Evidence packs What a clause needs you to be able to show

Mechanisms and evals

Mechanisms catalog Kill switches, appeals queues, and safe state controls Eval suites Governability suites and the cases inside them What we can check Every check mapped to the claim it establishes, and the claims nothing checks Validators Runtime assertions and middleware specifications Measurement tiers Tiered evidence, gaming patterns, and detection logic

Instruments

Delegation audit Bring a workflow: is the delegation still justified? System auditor Bring a prompt or architecture: where are the guardrails? Burden modeler Bring a workflow: who absorbs the friction, and how much? Record conformance Bring a record stream: does it earn the level it declares? All instruments Including the forecasters and simulators, which take numbers

Knowledge

Glossary Defined terms, each at a stable URL you can cite Capability taxonomy Domains, capabilities, and practices with owners and readiness Casebook Five public failures scored against the six state variables Field notes Working papers and empirical governance studies Theory Why the laws hold, kept apart from the requirements

Ecosystem & Community

GitHub Ethotechnics Studio Send feedback

Home
Validators

VAL-01

VAL-01 Burden Modeler — specification

The assertion, its inputs, and what a conforming implementation must return — not the tool. To model a workflow, open the Burden Modeler diagnostic →

Linked standard: STD-01

Estimated time to completion (minutes) Number of required steps Exit available on every screen?

Green Low risk: consent burden within bounds.

Fix this using MEC-04: The Hard Clock.

See also: Standards · Mechanisms

Ethotechnics Institute

An open framework for accountable AI systems — proposed standards, mechanisms, and diagnostics.

Content licensed CC BY 4.0

Method & Standards

The Method
Twelve Laws
Standards Register
Regulatory Crosswalks
Evidence Packs

Mechanisms & Diagnostics

Mechanisms Catalog
Diagnostics Suite
Delegation Audit
System Auditor
Evals & Coverage
Validators

Knowledge & Research

Glossary & Ontology
Capability Taxonomy
Failure Casebook
Theory & Foundations
Field Notes
Syllabus
Search

Institute & Governance

Start Here
About the Institute
Governance Process
Security Policy
Send Feedback
Reference API

Ecosystem & Connect

GitHub
Ethotechnics Studio
RSS Feed

Standards & Ontology Updates

Get notified when new standards, terms, or field notes are published.