# AI Learning & Knowledge Infrastructure
### A Proposal to the AI for Social Impact Alliance
---
## Why This Matters
Civil society organizations across Canada are facing the same challenge: AI is reshaping their landscape, but the conversation is dominated by two poles - corporate hype promising transformation, and blanket rejection treating all AI as inherently harmful. Neither serves communities well.
The organizations in this Alliance hold deep knowledge about community needs, social systems, and what "impact" actually means on the ground. But most lack the practical fluency to evaluate AI tools critically, deploy them on their own terms, or participate meaningfully in governance conversations about AI's future.
This isn't about "catching up" with the tech sector. It's about ensuring that the people closest to communities - the ones doing the actual work of social change - can make informed decisions about which tools serve their mission and which don't. It's about building the capacity to say "no" from a position of knowledge, not fear - and to say "yes" with clear eyes about trade-offs.
**The real risk isn't that civil society uses AI. It's that civil society gets left out of shaping it entirely.**
We've seen what happens when communities engage critically rather than reactively. People who understand both the limitations and the possibilities become the strongest advocates for responsible development, the sharpest critics of corporate overreach, and the most creative builders of alternatives.
This program creates that capacity.
---
## The Program
### Learning Events
A structured series of peer-driven sessions at three levels, delivered on open video platforms with recordings shared in a commons. Weekly drop-in coworking and one-on-one calls provide ongoing support between sessions.
- **Introductory** - Shared context-setting; surfacing what members already know and do. Honest discussion of concerns, harms, and trade-offs alongside opportunities.
- **Skills** - Practical, jargon-free sessions on digital knowledge portability, basic tooling (command line, code editors, AI assistants), and evaluating any tech stack for real ownership and control.
- **Technical deep-dives** - Co-designed with members; practitioner workshops on local model deployment, federated data, and open-source frameworks; knowledge mobilization labs for moving community knowledge into interoperable formats.
### Peer Support & Knowledge Mapping
- Weekly drop-in office hours and coworking sessions, hosted by Alliance member communities
- Scheduled one-on-one calls for focused problem-solving
- Direct access to contributor networks for deploying knowledge resources and open-source, privacy-preserving tooling
- Technical sensemaking for founders and leads: translating non-technically-stated operational challenges into specific, searchable language - and surfacing dissonance between stated needs and worker experience before it derails a project
- Mapping existing social tech projects across Canada's bioregions; connecting members to bioregional and domain-specific knowledge commons
---
## Budget & Resourcing
Target: **CAD $55,000-$115,000**, shared across two funding streams.
| Stream | Target |
|--------|--------|
| Participant co-funding | CAD $30,000-$50,000 |
| Values-aligned sponsorship | CAD $25,000-$50,000 |
**Co-funders** contribute on a pay-what-you-can basis (target: 3-5 organizations at CAD $10,000 each) and receive curriculum input, a dedicated privacy-preserving AI assistant for their own knowledge management, and additional one-on-one and group support.
**Values-aligned sponsors** (open-source technology foundations, public interest tech funders, and comparable networks) receive content input, live access to Alliance members, and free slots for Canada-based AI infrastructure projects.
Program resources are governed transparently through the Alliance's governance structure. All financial flows are documented on public ledgers.
**Labour** is contributed by a distributed network drawing from Alliance member organizations and aligned networks including Green Pill Network, Bread Cooperative, OpenCivics Labs, and others.
---
## What We're Asking For
1. **Co-funding commitments** - Even soft or conditional commitments give us a credible signal of demand when approaching sponsors.
2. **Permission to name the Alliance in sponsor outreach** - Any formal partnership agreements will come back to Community Governance for review before signing.
3. **Feedback and topic requests** - What questions are most pressing for your members? What domains are most represented in your network? Early input shapes the curriculum.
---
## Addressing Common Concerns
Based on community feedback, we want to be upfront about tensions this work navigates:
**"AI is environmentally destructive."** Yes, large-scale AI training has significant environmental costs. This program specifically prioritizes local models, efficient architectures, and community-owned infrastructure over corporate cloud services. We also build capacity to critically evaluate these trade-offs rather than accepting industry claims at face value.
**"AI displaces workers and steals from artists."** These are real harms happening now. This program doesn't shy away from them - in fact, building AI literacy is how communities develop the knowledge to advocate effectively for labour protections, IP reform, and accountability measures.
**"You can't 'shape' tools owned by billionaires."** Agreed - we can't reform OpenAI or Google from the outside. But we can build alternatives: open-source models, community-owned infrastructure, local deployments that keep data sovereign. That's a core focus of this program.
**"AI steals from artists and displaces workers."** Yes. AI models are trained on copyrighted work without consent, and automation is already displacing jobs across creative and knowledge-work sectors. This program includes these realities in its curriculum - not as abstract policy issues but as lived harms that shape how communities should evaluate and govern these tools. We advocate for consent-based training data, fair compensation, and labour protections as part of responsible AI engagement.
**"AI companies are complicit in military and surveillance operations."** This is documented fact - from Palantir partnerships to contracts with immigration enforcement and military agencies. This program doesn't pretend corporate AI is neutral. We build literacy that includes understanding these power structures, so communities can make informed decisions about which tools to use, which to reject, and which alternatives to build.
**"Civil society should reject AI entirely."** We respect this position but disagree. Blanket rejection doesn't stop AI from being deployed in communities - it just means communities have no voice in how it happens. We believe informed, critical engagement is more powerful than disengagement.