AI stewardship · metacognition · accountability

A stewardship layer for decisions AI cannot own.

M.I.N.D. is building an upstream human-governance practice to help institutions examine how AI-supported decisions are framed, authorized, challenged, monitored, and revised—before consequences are scaled.

Start a stewardship conversation Understand the layer

Emerging practice · Seeking bounded pilots and research partners · Not a certification body

Purpose

Evidence

Impacts

Human
accountability

Recourse · learning · revision

Human accountable

Evidence seeking

Life affirming

Globally interoperable

The institutional gap

Standards matter. Decisions still happen through people.

Policies and technical controls cannot, by themselves, determine whose knowledge counts, what level of harm is acceptable, or when an institution should refuse to automate.

M.I.N.D. focuses on the metacognitive and institutional conditions surrounding those choices. It is designed to complement—never replace—technical safety, security, law, risk management, domain expertise, and participation by people who may be affected.

Where M.I.N.D. sits

Between commitments and consequences.

A cross-system stewardship layer can help institutions translate values, standards, and obligations into documented decisions, conditions, monitoring, and recourse.

01Values, rights & law
Purpose, duties, dignity, ecological limits
02Governance & standards
Roles, controls, risk practices, evidence
03M.I.N.D. stewardship
Reflection, challenge, decision gates, learning
04AI use & operations
Design, procurement, deployment, oversight
05Outcomes & recourse
Monitoring, remedy, revision, withdrawal

Explore the stewardship layer

The working cycle

Pause before scale. Learn after action.

This proposed cycle is a practical structure for accountable reflection. It will be refined through bounded pilots, evidence, affected-party participation, and independent review.

Notice

Surface the assumptions, incentives, habits, and pressures shaping how an AI-supported decision is framed.

Map

Identify purpose, authority, affected communities, other life, ecological systems, dependencies, and uncertainty.

Deliberate

Bring evidence, lived experience, dissent, alternatives, and consequences into accountable human review.

Decide

Record who is responsible and whether to proceed, modify, constrain, pause, or stop—with reasons and recourse.

Learn

Monitor outcomes and incidents, listen for unanticipated effects, and revise or withdraw when evidence requires it.

See the full framework

The direction

Global in compatibility. Contextual in practice.

We are building toward a stewardship practice that can travel across institutions without erasing cultural knowledge, local conditions, or the voices of people most affected.

The aim is not centralized control. It is a shared discipline for responsible authority: interoperable with recognized governance frameworks, adaptable across settings, and accountable to human dignity, other life forms, ecological wellbeing, and future generations.

Read our stewardship commitments

Who we seek to work with

Institutions willing to learn in public-minded ways.

Early work should be bounded, documented, open to challenge, and willing to stop when the conditions for responsible participation are not present.

01Boards, executives, and governance teams
02Universities and research institutions
03Public-sector and civic institutions
04Foundations and civil-society organizations
05AI developers, deployers, and assurance partners

A bounded beginning

Bring the decision you are trying to govern.

A first conversation is for understanding the context, the people and systems affected, the responsibilities involved, and whether a M.I.N.D. pilot is appropriate.

Start a conversation