RegistryAbout the Registry

A public ledger for AI accountability

The Ethos AI Registry is a civic infrastructure project, not a company, government agency, or open-source hobby. It catalogs deployed AI systems, publishes NIST-aligned transparency scores, and documents community-reported incidents so that the public can understand and challenge automated decision-making.

Nonprofit initiativePublic-first dataIndependent of vendors
Mission

Building durable public trust in automated systems

Ethos AI Registry documents how consequential AI systems are deployed in public life and publishes the evidence needed to hold them accountable.

Our work is guided by NIST-aligned standards, independent review, and open methodology.

Our Vision

A public record for the algorithms that shape public life

We envision a future where every automated system that affects housing, benefits, employment, or civil rights is traceable, reviewable, and answerable to the communities it serves.

Why This Registry Exists

Transparency is the precondition for accountability

Most deployed AI systems are invisible to the people they affect. Without a shared public record, auditing, appeal, and reform are impossible.

The registry closes that gap with structured disclosures, independent scoring, and durable public evidence.

Miranda Bouldin, Founder and Executive Director of Ethos AI Registry

Leadership

Miranda Bouldin

Founder & Executive Director

Entrepreneur, government technology executive, and founder of the Algorithmic Accountability Project. With more than two decades of experience leading technology, logistics, workforce development, and AI initiatives, Miranda created Ethos AI Registry to strengthen transparency and public trust in automated systems.

Professional Credentials

  • Founder, Algorithmic Accountability Project
  • CEO, LogiCore Corporation
  • Founder, TechGeekette Foundation
  • Government Technology Executive
  • AI Governance Advocate

Connect & Learn More

Trust in artificial intelligence cannot be assumed—it must be earned through transparency, accountability, and independent oversight. Ethos AI Registry exists to help make that trust measurable.
Miranda Bouldin, Founder & Executive Director
Our Principles

Four commitments that guide every record we publish

Transparency

Open, understandable AI governance.

Independence

Objective assessments without commercial influence.

Accountability

Clear standards and measurable outcomes.

Public Trust

Technology that serves people responsibly.

Governance

Institutional stewardship

Ethos AI Registry is an initiative of the Algorithmic Accountability Project and is supported by the TechGeekette Foundation, a US 501(c)(3) public charity.

The registry is guided by transparent methodologies, evolving standards, and independent review processes designed to outlast any single administration, funder, or contributor.

Governance Structure

  • Advisory CouncilForming 2026
  • Technical Review BoardForming 2026
  • Community ContributorsOpen participation
  • Research PartnersBy invitation

Council and Board appointments, conflicts of interest, and funding sources are published annually.

Institutional context

About the Algorithmic Accountability Project

Ethos AI Registry is the flagship initiative of the Algorithmic Accountability Project, a public-interest effort dedicated to advancing transparency, accountability, and responsible AI governance through research, education, and civic technology.

The Project is supported by the TechGeekette Foundation, a US 501(c)(3) public charity that stewards its civic technology and education programs.

Learn more about the Algorithmic Accountability Project

Our ecosystem

TechGeekette Foundation
            │
            ▼
Algorithmic Accountability Project
            │
            ▼
      Ethos AI Registry
TechGeekette Foundation
The nonprofit organization (techgeekettefoundation.org).
Algorithmic Accountability Project
The research and public-policy initiative.
Ethos AI Registry
The public platform for AI transparency, accountability, and governance.
Purpose

Why the registry exists

Consequential AI systems are already deciding who gets a loan, which neighborhoods are policed, who receives benefits, and who is hired. Most of those systems are invisible to the people they affect.

Systems are hard to find

Agencies and vendors rarely publish a complete list of the AI tools they use. The registry gathers public records, operator disclosures, and reported harms into one searchable place.

Harms repeat without accountability

When a biased system causes harm, the failure is often buried in a procurement contract or dismissed as a technical glitch. Public records make patterns visible and repeatable corrections possible.

Procurement lacks shared standards

Buyers, auditors, and advocates rarely use the same vocabulary. The registry applies a NIST-aligned scoring rubric so that comparisons are consistent and evidence-based.

Impact

Why AI transparency matters

Transparency is not a compliance checkbox. It is the precondition for every other form of accountability: auditing, appeal, journalism, and public oversight.

Fairer due process

When people know a system is being used, they can challenge decisions, request records, and demand human review.

Better engineering

Public scrutiny surfaces failure modes early, before they scale into systemic harm.

Smarter procurement

Agencies and vendors can compare systems against a common baseline instead of marketing claims.

Informed public debate

Journalists, researchers, and communities can ground policy debates in specific, verifiable cases.

Stewardship

Who operates the registry

The registry is a program of the Algorithmic Accountability Project, which is an initiative of the TechGeekette Foundation, a US 501(c)(3) public charity.

That means we are governed by a board, file a public Form 990, do not sell ratings or placement, and cannot be acquired by a vendor or government agency. Our mission is encoded in our charter: equitable technology, civic transparency, and algorithmic accountability.

501(c)(3) tax-exempt public charity
Independent of vendors, ratings markets, and political campaigns
Funding and conflicts of interest published annually

Legal status

TechGeekette Foundation

A 501(c)(3) public charity

EIN

93-4079110

Program

Algorithmic Accountability Project

Participation

How organizations participate

Operators, agencies, and vendors can register systems, disclose controls, and respond to incidents. Participation is public, documented, and auditable.

1

Register the system

Submit the operator, system name, intended use, deployment context, and public reference materials.

2

Complete the rubric

Answer a NIST-aligned questionnaire covering governance and transparency controls.

3

Publish the score

Receive a public Transparency Score, sources, and reviewer notes. Scores are re-reviewed annually or on material change.

4

Respond to incidents

Operators are notified before adverse records are published and may file corrections or appeals.

Community reporting

Who can submit incidents

The registry is built on the principle that the people most affected by a system often know the most about its failures.

We accept reports from

  • System operators and vendors with direct knowledge
  • Public servants, auditors, and procurement officers
  • Researchers, journalists, and civil-society organizations
  • Community members and affected individuals

Anonymous reports are accepted

We know that retaliation is a real risk for whistleblowers and affected communities. Anonymous submissions are held to a higher corroboration bar but are not excluded.

Every published incident links to a source, reviewer, and operator response window. The record belongs to the public.

Trust

Why trust the registry

Trust is earned through transparency of process, not authority. We publish how we score, how we review, how we correct mistakes, and who pays the bills.

Open methodology

Every score is derived from a published rubric. The full crosswalk, point values, and version history are on the Methodology page.

Read the methodology

Named reviewers

Every record is reviewed by a named human reviewer who discloses conflicts of interest. No reviewer evaluates a system they are employed by or financially tied to.

Fair appeals

Operators and affected parties can appeal adverse records within 30 days. Corrections are appended to the original record; retractions are publicly noted.

Frequently asked questions

Is the registry a government database?

No. The registry is an independent nonprofit civic project. We ingest public government records, but we are not operated by a government agency.

Is the registry open source?

The methodology, data model, and scoring rubric are public. The registry is a service operated by a nonprofit; raw data is published under an open Creative Commons license.

Does the registry sell ratings or placement?

No. We do not accept payment for higher scores, preferred placement, or removal of adverse records. Funding comes from grants and public donations.

Can a vendor dispute its score?

Yes. Every operator receives notice before adverse publication and can file an appeal. The full appeal history is part of the public record.

What organizations are eligible to register a system?

Any organization that deploys or operates an AI system in a public-facing or high-stakes context may register. This includes government agencies, vendors, non-profits, research institutions, and private companies whose systems affect public services, employment, lending, criminal justice, health, education, or benefits.

Does it cost money to register or participate?

No. Registration, scoring, and basic incident response are free. The registry is a nonprofit public service; we do not charge operators for listing, scoring, or corrections. Optional donations and grants fund the infrastructure, not fees from participants.

What happens after an organization registers a system?

After registration, a Community Reviewer validates the submission for completeness and verifiable sources. The organization then completes the NIST-aligned transparency questionnaire. Once reviewed, the system receives a public profile with a Transparency Score, review notes, and links to source materials. The profile is re-reviewed annually or whenever a material change is disclosed.

What if an organization cannot answer every rubric question?

Partial answers are accepted and scored as partial credit. The registry is designed to reveal transparency gaps, not punish them. Operators can update their responses as their governance practices mature, and the score is recalculated on the next review cycle.

Help make AI systems visible

Whether you operate a system, witnessed a harm, or simply want to understand what is deployed in your community, you have a place here.