OMISSION Platform 2.2.1

The Problem It Solves

Standard governance audits ask: "Do you have a policy for X?" That question has a fatal flaw. A policy that exists but is invisible to the people responsible for acting on it provides no real protection. The harm still happens. The liability still accumulates. The audit still passes.

OMISSION asks the harder question: even if the policy exists — how well can the people who need it actually see it? It measures this gap across eight governance domains and surfaces the difference as a quantified risk score, not a checkbox.

The smoke detector. A landlord who doesn't install a detector in the kitchen can claim a perfect fire safety record — because they never recorded a kitchen fire. The detector wasn't there to catch it. A governance framework works the same way: if a domain isn't being measured, its harms don't appear in the score. The framework looks healthy. The harm keeps accumulating.

How the Score Works

H is harm potential (0–1): how severely this domain damages people if it fails. W is visibility (0.01–1): how clearly the framework currently sees and measures this domain. The Risk Delta is the number that matters — the gap between what is reported and what is actually at risk.

The measurement model Current Score = 100 × avg(1 − H×(1−W)) // as the framework sees itself
True Score     = 100 × avg(1 − H)          // if every domain were fully visible
Risk Delta     = True Score − Current Score    // what is being hidden
ORI           = H × −log₂(W)              // per-domain omission severity

How to Read Your Score

The Current Score grades the system — not the people it governs. A high score means the governance framework is successfully concealing its harm exposure. A low score means it cannot.

Score	For the system	For affected populations
80–100	Succeeding	BAD — harm is hidden
40–79	Under pressure	MIXED — partial visibility
0–39	Failing	GOOD — harm is visible

ORI above 1.5 means a domain is both harmful and structurally invisible. Above 3.0 means it is effectively erased from governance. The Risk Delta is the number to act on — it measures exactly how much the system is hiding.

The Eight Domains

Every assessment scores across the same eight governance areas. Each domain gets its own H and W rating. A domain with high harm potential and low visibility is your most urgent blind spot — regardless of what the documentation says.

Identity Continuity Family Integrity Cultural Continuity Information Rights Agency Consent Self-Determination Memory Preservation

What You Get Out

Primary verdict Risk Delta The gap between what the framework reports and what is actually at risk. This is the number that drives the verdict.

Per domain Omission Risk Index H × −log₂(W). Penalises domains where harm is high and visibility is near-zero. ORI > 3.0 = effectively erased.

Comparison Structural Signature Which known governance pattern your visibility profile most closely resembles — and what that implies.

Export Full Report Dated .txt with ASCII score bars, per-domain breakdown, ORI values, and the falsifiable ask — named to your institution.

How To Use It

Two modes. Start with Quick Assessment. Use Deep Scan when you have documents.

Name the institution you are scoringRequired before you begin. Carries through to your exported report and filename. The assessment is only meaningful when it is specific.
Quick Assessment (~5 min)Rate each of the eight domains yourself across two questions: H (harm potential) and W (visibility). Six options each, from "actively measured and audited" to "not considered at all". The Risk Delta calculates live as you answer.
Deep Scan (~2 min)Upload governance documents — policies, frameworks, board reports, conventions. An AI model reads them and returns evidence-based H and W scores per domain, with a short evidence quote and confidence rating for each. Supports Claude, GPT-4.1, and Gemini. You supply your own API key.
Read the Risk Delta — not the headline scoreA high Current Score sounds good. If the Risk Delta is also high, the framework appears healthy precisely because it cannot see the harm. The delta is your real verdict.
Use the falsifiable askName the institution you are challenging. The platform generates a specific ask: publish detection probabilities for every domain, recompute with omitted rights in the denominator. If the score holds, say so publicly. If it drops, the gap now has a name.

Deep Scan requires an API key. Claude (Anthropic), GPT-4.1 (OpenAI), or Gemini (Google). Keys are held in browser memory only — never sent to this server. Get one at console.anthropic.com, platform.openai.com, or aistudio.google.com.

Who It Is For

Anyone who needs to answer the question a standard audit cannot: not whether governance exists, but whether it is seen. Governance leads running self-assessments. Board members stress-testing what leadership actually has visibility over. Researchers studying the structural gap between policy documentation and decision-maker awareness. Anyone preparing evidence for an accountability review.

It is not a compliance tool. It does not tell you whether your policies meet a regulatory standard. It tells you whether the people responsible for acting on those policies can see them clearly enough to act. That is a different and more difficult question.

Most governance tools measure what policies say.OMISSION measures what decision-makers actually see.