Early Design Phase · March 2026

Community knowledge is power.
Only if communities keep it.

The Civic Knowledge Mesh is provenance-anchored infrastructure that lets communities custody their own knowledge, collaborate with AI on their own terms, and never hand control to a platform they don't own.

Get in touch →

Two crises are converging

Communities sitting on deep, verified, human-generated knowledge are about to be in a position of genuine power — for the first time. But only if the infrastructure exists to help them act on it.

Platform capture

Communities generating irreplaceable knowledge — Indigenous land stewardship, climate adaptation strategies, community health intelligence — must hand it to platforms they don't own, on terms they didn't set. Value accumulates on the platform. Community stewardship is replaced by dependency.

AI is eating its own tail

AI systems are running out of quality data. The web is flooding with AI-generated content that degrades model quality when it re-enters training pipelines. The information commons is being quietly poisoned. Authentic, human-origin knowledge is becoming scarce — and valuable.

What we're building

CKM is not a new platform. It is a layer that sits beneath the tools communities already use — adding custody, consent, and a verifiable chain of authenticity to every piece of knowledge they create.

Local custody
Nothing is uploaded to a central repository. AI capability comes to the data — not the data to a cloud.
Tamper-evident provenance
Origin, authorship, version history, and licensing terms travel with every knowledge artifact. This record is the truth in the system.
Encoded consent
Rights and licensing terms are machine-readable and embedded in the protocol. The system enforces them.
Federated AI
Local AI runs on modest hardware. Across the network, models improve from patterns across many nodes without any raw data leaving home.
Uncapturable governance
No single actor — corporate, governmental, or individual — can capture the mesh without the network detecting it.

The provenance advantage

CKM's provenance record is, structurally, a certificate of human origin. As AI-generated content degrades the open web, verified human knowledge becomes the premium resource for the next generation of AI training. CKM gives communities the infrastructure to license that knowledge on their own terms — attributed, compensated, governed by the community that produced it.

How it works

A concrete example. No abstractions.

01

A community creates knowledge

A climate adaptation network documents a flood mitigation strategy that worked in their region — field observations, local data, community input. Real knowledge from real experience.

02

CKM stamps it with provenance

Who wrote it. When. What sources they drew on. What terms govern its use. This record is cryptographically signed and travels with the knowledge everywhere it goes. It can't be stripped out.

03

The knowledge stays home

Nothing is uploaded to a central cloud. The community's AI tools run locally, drawing on their own knowledge base. Other nodes on the mesh can discover that the knowledge exists — but not access it without permission.

04

An AI company comes knocking

They need high-quality, verified, human-origin training data. The community sees the request, reviews the terms, and decides: yes, no, or yes-but-only-for-this-purpose. The provenance record enforces whatever they choose. They get attributed. They get compensated.

Who we serve

Starting narrow. Research groups first — they have the best data, the deepest incentive for provenance, and the most aligned funding model.

Tier Audience Why now
Tier 1 Indigenous communities & First Nations data governance $73.5M federal mandate for Regional Information Governance Centres. OCAP® principles map directly onto CKM's architecture.
Tier 1 Academic research consortia EU INFRAEOSC €200M+; NFDI €250M+/yr. Funders now require FAIR data + W3C PROV compliance. No affordable federated solution exists.
Tier 2 Community health networks Rural programs can't share knowledge across clinics. Local-first, offline-capable architecture is a direct fit.
Tier 2 Climate adaptation networks Cross-regional coordination with poor data sharing. CanAdapt is our first pilot with direct access through existing relationships.

Team

Architecture and complexity science — the combination is intentional.

Jane Zhang

Architecture & Prototyping

Research, architecture, prototyping. Former ED at TechSoup Canada, Chief Digital Officer at Centre for Social Innovation. CanAdapt background gives direct access to the first pilot and deep understanding of the problem.

Kirsten Wright

Strategy & Partnerships

Managing Director, WICI (University of Waterloo). Complexity science, interdisciplinary research, systemic transformation and resilience.

Business model

Grants anchor early development. Earned revenue builds sustainability.

Now
SSHRC partnership grant, Sovereign Tech Fund, Ford/Mozilla public interest AI, INFRAEOSC, MITACS Accelerate
Year 2+
Hosted node services, technical integration, training and capacity building
Year 3+
Consortium membership, provenance-verified training data licensing — high-margin, no incumbent

The window is open.

The technical foundation is ready. The AI data crisis is creating urgency. The community is forming. The moment to build this is now.

[email protected]