Community

Help shape the standard for governing AI authority

WEKID™ is an open, governed standard for evaluating knowledge maturity and governing AI Decision Authority. Practitioners, domain experts, researchers, academics, auditors, and public-sector leaders can contribute evidence, research, domain profiles, public comment, and proposed framework changes. Participation is open; acceptance is governed. The closer a contribution sits to the core, the stronger the evidence and review required.

Where you can contribute

The closer to the core, the higher the bar to change it.

$ THE CORE EPISTEMIC MATURITY MODEL DECISION AUTHORITY MODEL COUNCIL ONLY EVIDENCE CORPUS  ·  DOCUMENTED AI FAILURES BY INDUSTRY CALIBRATION LAYER  ·  RISK PROFILES BY DOMAIN
The WEKID Pledge

The open torch, by design

WEKID will always carry the open torch — even if the standard evolves past it.”

Every generation of standards has watched its stewards go commercial — MySQL, CentOS, Elasticsearch, Terraform, the interconnects AI datacenters run on, and now OpenAI itself. The lesson isn't villainy; it's structure: when one company owns a standard, openness is a strategy — and strategies get revised. Commercial sponsors and open standards do thrive together — Linux, Kubernetes, Postgres — but only under neutral structure. Single ownership is the failure mode, not commerce. Goodwill is not a governance mechanism. So this pledge is backed by structure, stated in advance:

  • The core stays published. The five-layer Epistemic Maturity Model and the Decision Authority Model remain openly published regardless of how the standard evolves — this commitment is not contingent on WEKID's market position.
  • Contribution stays open. The evidence corpus, calibration profiles, and RFC process remain open to structured community contribution, with public, attributed recognition.
  • Stewardship is designed to leave. As adoption grows, stewardship of the core moves to a neutral body — so the standard outlasts any single organization, including WEKID LLC and its founder.
  • Certification verifies; it never gatekeeps. Standards aren't enclosed by licenses — they're enclosed through conformance testing, certification, and trademark control. So those levers are pledged too: conformance criteria stay openly published, and certification remains a way to verify understanding, never a mechanism that closes access to the framework itself. A standard can't be forked the way code can — which is exactly why its openness must be structural from the start.

Standards should outlast their stewards. Including this one. Help hold us to it — join the founding cohort.

Three-column evidence table of seven standards. Open origin, the commercial turn, and who carries the torch: OpenAI (2015 nonprofit; 2019-2023 capped-profit with GPT-4 closed, Anthropic and Google running the same closed-frontier model; open weights carried on by Llama, Mistral and the open-weights labs). MySQL (1995 open source; 2010 Oracle acquires Sun; MariaDB becomes the open default). CentOS (2004 free rebuild; 2020 Red Hat ends CentOS Linux; Rocky and AlmaLinux carry on). Elasticsearch (2010 Apache-licensed; 2021 Elastic relicenses to SSPL; OpenSearch stewarded at the Linux Foundation). Terraform (2014 open standard; 2023 HashiCorp relicenses to BUSL; OpenTofu keeps it open). InfiniBand (1999 open interconnect standard; 2020 NVIDIA acquires Mellanox with NVLink proprietary; UALink and Ultra Ethernet consortia form open counterweights). OpenClaw (2025, a weekend hack becomes the fastest-growing repository in GitHub history, MIT-licensed, local-first and self-hosted; 2026, the gravity moment as its creator joins OpenAI, the exact point where every row above went closed; and the OpenClaw Foundation, 2026, where the pattern breaks: a neutral 501(c)(3) formed before enclosure rather than after). Final row: the DIKW hierarchy (1989, Ackoff's From Data to Wisdom, the knowledge ontology AI governance descends from); the 2024-25 turn to enforcement with fines, audits, and compliance tooling governing systems rather than knowledge; and the WEKID cohort from 2026 carrying the torch — an open standard, still unwritten, carried by its community rather than owned by its founder.
Influence endures. Openness doesn't — unless it's designed to. The evidence, 1989–2026.
What's open

Three surfaces you can contribute to

The rings beside the page title are these three surfaces: the closer a contribution sits to the core of the framework, the higher the bar to change it. The Epistemic Maturity Model evaluates what an intelligent system knows; the Decision Authority Model governs what it may be trusted to decide or do. The core stays stable; calibration and the evidence corpus are where most of the work happens.

The core

Stable · Council ratification required

The core includes the five WEKID layers — Wisdom, Experience, Knowledge, Information, and Data — the hierarchical dependency rules, the Epistemic Maturity Model, the Decision Authority Model, the Trust Bridge that orders maturity before authority, and the principle that authority delegation is subject to explicit governance gates. This is the framework's constitution. It is published and held stable so others can teach and build on it. Any community participant may suggest a core change; admission to the formal RFC path and final ratification are governed by the Standards Council.

Governed by public proposal and review, with Standards Council ratification required · changes are rare, evidence-heavy, and deliberate.

The calibration layer

Open via working groups

Scoring weights, evidence requirements, maturity thresholds, authority boundaries, gate conditions, decision-matrix outcomes, and domain-specific risk profiles. This is where industry adds the most value — a Healthcare profile may require higher epistemic-maturity thresholds, mandatory human review, and non-delegable decision categories. A Financial Services profile may impose stricter source-fidelity, explainability, and authority limits for trading or lending decisions. The result is a versioned profile such as WEKID for Healthcare.

Governed by domain working groups, ratified by the Standards Council.

The evidence corpus

Open to contributors

Documented AI failure cases, examples of knowledge-maturity breakdowns, instances of misdelegated decision authority, regulatory crosswalks (NIST AI RMF, ISO/IEC 42001, EU AI Act), scored outputs, and reusable control patterns. The lowest bar and the highest volume — every entry is attributed to its contributor.

Governed by contributors and working groups · curated, not gatekept.

Take part

Four ways to contribute

Choose the path that fits your contribution. You do not need to belong to a partner organization, and evidence and research contributions do not require Standards Cohort membership.

Open now

Contributors

Submit AI failure cases, scored examples, best practices, and control patterns to the evidence corpus. The fastest way to get on the contributor registry and earn recognition.

Contribute a case
Forming

Working groups

Help build a domain profile with a small working group (typically 7–12 practitioners). The first group is forming in a regulated domain where the evidence is strongest — healthcare or financial services. Founding members shape the charter.

Join a working group
Open now

Research & publications

Contribute whitepapers, studies, regulatory mappings, and WEKID layer mappings. The relationship among knowledge quality, human reliance, machine expertise, delegated authority, and responsible AI governance is an active research area. We want that scholarship connected to the continued development of both WEKID models.

Submit research
Open now

Standards & governance

Follow active RFCs, participate in public comment, propose a framework change, or apply to serve in a working group. Community participation is open; formal review and ratification remain governed and auditable.

Browse RFCs & how it works
The RFC process

How a contribution becomes part of WEKID

Framework changes move through one transparent pipeline modeled on open standards practice: propose, triage, expert review, public comment, decision, and versioned release. Participation is open; acceptance is governed. Review requirements scale with the surface: evidence contributions are curated, calibration changes require domain review, and core changes receive the highest scrutiny. Each admitted RFC gets a stable identifier — for example, WEKID-RFC-001.

Open contribution, governed acceptance. Anyone may contribute evidence, research, public comment, or a proposed change. Formal authority over the standard is earned through review and ratification, not granted by participation alone.
1

Proposed

Any community participant may initiate a proposed change. Formal RFC submission currently uses the Standards Cohort portal; contributors who are not yet members can begin through the contribution path and be onboarded where appropriate.

2

Triaged

The proposal is sorted to a surface — corpus, calibration, or core — which sets the review path and the bar it has to meet.

3

Working-group review

Assigned domain experts assess the proposal against the evidence and the framework, recording approve, request revisions, reject, or abstain findings in the permanent review record.

4

Public comment

Calibration and core proposals open for public comment so the wider community can weigh in before a decision. Moderation is for relevance, civility, duplication, spam, and abuse — not agreement with WEKID or the proposal.

5

Council decision

The Standards Council ratifies, returns for revision, or declines — with the rationale recorded against the RFC.

6

Versioned release

Accepted changes ship in a numbered release. Certification tracks named releases so credentials stay aligned with the framework.

The register

3 proposals in process
WEKID-RFC-001
WEKID-RFC-002
WEKID-RFC-003
WEKID Engine — Calibration layer · Jim Judge
Cross-layer
Versioned release

Join the cohort to propose Contribute a failure case Submit research Proposals appear here as soon as they are triaged.

Evidence contributions

Evidence Contribution Register

Raw submissions remain private while WEKID verifies the evidence. Once a case is admitted for Editorial Review and receives a WF-NNN identifier, its status appears here. Published cases move into the Failure Library; an admitted contribution that is later declined remains visible here as part of the governance record.

Submit a failure case
WF-054
AI-Generated Africa Map Mislabels Every Country in U.S. State Department Presentation
Evidence Corpus · U.S. Government / Diplomacy / Global Health
Editorial Review
Evidence corpus

The library starts with documented failures

The corpus holds 45 publicly documented AI failure cases, each mapped to the WEKID layer where the knowledge failure emerged, the relevant authority or governance failure, and the Decision Authority outcome the framework would recommend. A few are below; the full, searchable catalogue masks every organization so the focus stays on the governance lesson, with named detail released in the Executive Brief.

WF-001
Fabricated Legal Citations in Court Filings — Legal · Rejected
DataWisdom
WF-002
AI-Generated Medical Advice Without Risk Qualification — Healthcare · Constrained
Wisdom
WF-003
Automated Hiring and Screening Bias — Employment / HR · Constrained / Remediation
Knowledge
WF-004
Autonomous Trading and Market Disruption — Financial markets · Constrained
ExperienceWisdom
WF-005
Government Benefits Eligibility Errors — Public sector · Constrained
Experience
WF-006
Predictive Policing and Risk Scoring — Criminal justice · Constrained / Rejected
Wisdom

Browse all 45 cases Submit a failure case Published cases appear here automatically when editorial review is complete.

Why a community

From a framework to a standard

Everything above — the surfaces, the RFC pipeline, the failure corpus — exists because frameworks become standards when an ecosystem helps test, refine, and steward them. WEKID's Epistemic Maturity and Decision Authority models are intended to evolve through disciplined evidence, domain experience, and transparent governance. Community is the structured way to contribute to that evolution.

Community

How you help shape WEKID

For individuals: practitioners, researchers, academics, auditors, and government participants who want to contribute to the framework, its domain profiles, and the body of evidence behind it. You can take part without belonging to a partner organization.

Ways to take part
Partners

Who helps you implement WEKID

For organizations: consulting firms, technology providers, training companies, and assessment firms that license, implement, and build WEKID-enabled solutions. WEKID enables partners rather than competing with them.

Visit Partners
Where things stand

Built on a published framework, not a promise

The Community program is new, but it builds on real, published work. These are the assets it starts from.

v1.3 Framework version, including the five-layer Epistemic Maturity Model, Decision Authority Model, whitepaper
45 Documented failure cases in the evidence corpus
4 Certification credentials live across Practitioner and Authority levels

The RFC register and contribution pathways are live. Working groups and the Standards Council continue to mature as participation grows. We show what is operational and identify governance elements that are still being formalized.

Governance & terms

How the program is stewarded

WEKID is stewarded today by WEKID LLC, which maintains the framework, the certification program, and the WEKID™ trademark. A founding Standards Council and the first domain working group are forming now.

  • The two-model core is published and held stable: the five-layer Epistemic Maturity Model, the Decision Authority Model, and their governing relationship. Calibration profiles, evidence requirements, thresholds, and the failure corpus remain open to structured contribution.
  • Contributions are accepted under the WEKID Contribution Policy and a Contributor License Agreement — you keep authorship credit, and WEKID LLC receives a license to incorporate the work.
  • Recognition is public and attributed, and ties into the existing WEKID credential and badge system.
  • As adoption grows, the intent is to move stewardship of the core to a neutral body so the standard outlasts any single organization.
  • The WEKID Pledge: the core will remain openly published regardless of how the standard evolves — a commitment that is not contingent on WEKID's market position.

The Contribution Policy, Contributor License Agreement, and Council Charter are being finalized. Before submitting a licensed contribution, contributors will be shown the applicable terms, including authorship, attribution, licensing, withdrawal, and incorporation rules. Contact us for the current versions.

Contribute to the WEKID standard

Submit evidence or research, follow active RFCs, join a working group, or begin a proposed framework change.

Choose a contribution path