Skip to main content

About Governance Commons

Capability should never silently become authority.

Governance Commons is a maintainer-led public project for making governance of human–AI and multi-agent systems more inspectable, teachable, and contestable.

Definition

A commons for governed human–AI systems.

Governance Commons brings together specifications, evidence contracts, reference tooling, research, education, and public explanation. Its subject is sociotechnical: not only what systems can do, but who may authorize them, how decisions are reviewed, what evidence survives, who is affected, and how people can refuse, appeal, override, or revoke delegated action.

“Commons” describes the intended public purpose and direction. It does not claim that GC is already a neutral institution, a formal standards body, or a community governed by broad participation.

Mission

Make authority, evidence, and recourse visible.

Make governance for human–AI systems inspectable, teachable, and contestable—so capability never silently becomes authority.

GC works toward systems in which another person can determine who authorized an action, what boundaries applied, what evidence was retained, who remains accountable, and what can happen when the action is wrong.

Why I started it

I began with questions, not credentials.

I did not begin Governance Commons as a software engineer, web designer, computer scientist, or governance professional. I began as a person trying to understand a practical gap: increasingly capable AI systems can act, coordinate, and produce evidence, but capability is not the same as legitimate authority.

The project grew through persistent questions. Who allowed this action? What was delegated? What happens at a handoff? What evidence remains afterward? Who can object? Who can stop it? Who is accountable when technical capability crosses a human or institutional boundary?

I use AI-assisted research, drafting, prototyping, and technical implementation to explore those questions in public. I direct the scope, make the project decisions, test claims, preserve boundaries, and revise the work when the evidence does not support what I hoped was true. The method is imperfect and still developing; making it inspectable is part of the point.

Goals

Build useful common ground without claiming the whole field.

Success is not a large umbrella claim. It is a set of narrower outcomes that other people can inspect, question, implement, and improve.

Clarify

Separate capability, authority, delegation, evidence, accountability, and recourse instead of allowing them to blur together.

Publish

Create versioned specifications, evidence contracts, tools, and explanations with visible maturity and provenance.

Connect

Relate technical mechanisms to institutions, power, trust, legitimacy, and the people affected by delegated machine action.

Teach

Make difficult governance questions approachable to builders, reviewers, organizations, learners, and affected communities.

Invite critique

Make assumptions and limitations visible enough that others can challenge, test, adapt, or reject them.

Earn stewardship

Work toward broader participation and durable stewardship without claiming neutrality or community authority before they exist.

Working method

AI-assisted, maintainer-directed, evidence-conscious.

AI systems help research, draft, critique, prototype, implement, and test GC materials. They do not become the source of authority merely by producing fluent work. Claims are scoped to the evidence available, sources retain their provenance, and public status distinguishes released specifications from candidates, emerging work, tooling, and maintainer coordination.

The project is revised when validation, deployment, compatibility, or outside evidence contradicts an earlier assumption. Source code, publication, runtime behavior, adoption, and institutional legitimacy are treated as different kinds of proof.

Participate

Bring questions, criticism, experience, and craft.

Governance Commons needs more than code. Useful contributions include challenging an assumption, testing a specification, reporting confusing language, sharing an implementation experience, reviewing evidence, improving accessibility, contributing research, strengthening a visual explanation, or identifying who is missing from the conversation.

You do not need a formal title or technical credential to participate. State what you observed, what you think should change, and what experience or evidence informs your view. Disagreement is welcome when it is specific, respectful, and open to examination.

Offer feedback

Report something unclear, incomplete, inaccessible, or unsupported on the public website.

Website issues →

Collaborate

Start with the problem, affected people, intended outcome, and the role you would like to play. Early collaboration should make authority and expectations explicit.

Explore the public work →

Support the work

Help sustain careful public work.

Financial support could help cover hosting, research materials, accessibility, design, validation, independent review, and the time required to maintain public specifications and tools.

Funding and donation options are coming soon. Verified support links will appear here as they become available, together with clear information about the payment route, use of funds, fees, tax treatment, and the relationship between financial support and project authority.

Financial support will not purchase certification, favorable findings, control over specifications, or greater authority in governance decisions.

Limits

A public project, not borrowed authority.

Governance Commons is not a regulator, certification authority, accredited standards body, legal adviser, or substitute for the authority of the organizations and communities whose work it references.

Current boundary: RFC-0001 is accepted, but its Public Boundary Model remains inactive until activation gates G-1 through G-5 verify. GC v1.2.0 remains the pinned public baseline during that transition.