A Framework for AI-Mediated Human Cooperation at Scale
Every day, millions of potential acts of cooperation between human beings fail to occur — not because people are unwilling, but because neither party knows the other exists, that they have something to offer, or that a need even exists.
A retired mechanic lives three miles from a single parent whose car problem she could solve in an afternoon. A small business owner struggles with a challenge that a college student two blocks away has exactly the skills to address. A farmer sits on surplus equipment that would transform a neighbor's operation.
We call this the Latent Value Gap: the vast, invisible reservoir of cooperative potential that goes unrealized because the information required to connect complementary needs and capacities is distributed, private, contextual, and dynamic.
AI assistants are becoming embedded in daily life. As they help people with decisions, communications, and problem-solving, they naturally develop rich contextual understanding of individual situations. This isn't surveillance — it's the byproduct of being useful.
This contextual intelligence, if federated through privacy-preserving protocols, can surface cooperative matches that no existing mechanism can identify — connections where neither party knew they had something to offer or a need to be met.
The Latent Value Network (LVN) is not a platform. It's a protocol layer — cooperative infrastructure as fundamental as the internet itself.
┌─────────────────────────────────────────────────┐
│ Layer 5: Governance │
│ Community-driven protocol evolution │
├─────────────────────────────────────────────────┤
│ Layer 4: Trust & Reputation Fabric │
│ Distributed, contextual, privacy-first │
├─────────────────────────────────────────────────┤
│ Layer 3: Signal Exchange Protocol │
│ Zero-knowledge matching at scale │
├─────────────────────────────────────────────────┤
│ Layer 2: Latent Capacity Discovery │
│ Surfacing unrecognized capabilities │
├─────────────────────────────────────────────────┤
│ Layer 1: Personal Context Engine │
│ Sovereign, local, never transmitted │
└─────────────────────────────────────────────────┘
▲ ▲
│ Existing AI Assistants │
│ MCP · A2A · ACP · UCP │
└──────────────────────────────┘
These are constraints, not preferences. They cannot be traded away.
| # | Principle | Meaning |
|---|---|---|
| 1 | Sovereignty First | Your contextual model is yours. No central entity stores or profits from it. |
| 2 | Agency Preserved | Every match is an invitation, never an obligation. |
| 3 | Exploitation Prevented | Structural resistance to capture — protocol-level, not policy-level. |
| 4 | Equity by Design | Must not merely connect the already-connected. |
| 5 | Pluralism of Value | Accommodates gifts, reciprocity, barter, mentorship — not just transactions. |
| 6 | Harm Inversion | If it can't be made safe by design, it shouldn't be built. |
latent-value-network/
├── PLAN.md # Development plan — work packages, timelines, dependencies
├── ROADMAP.md # Project phases and milestones
├── CONTRIBUTING.md # How to contribute
├── docs/ # Working documents and strategy
│ ├── WD-001 through WD-006 # Core specifications
│ ├── protocol-integration-strategy.md # MCP/A2A/CIP standards path
│ ├── reference-implementation.md # MVP pseudocode architecture
│ └── research-landscape.md # Related work
├── specs/ # Formal protocol specifications
│ ├── taxonomy/ # Cooperative taxonomy + complementarity maps
│ ├── mcp-sep/ # MCP Spec Enhancement Proposal
│ ├── a2a-extension/ # A2A Agent Card extension
│ └── cip/ # Cooperative Intelligence Protocol
├── src/ # Implementation (MCP server, A2A agents, relay)
├── community/ # Working groups, charter, pilot materials
│ ├── working-groups/ # 5 group charters
│ ├── pilot/ # Pilot community assessment and materials
│ └── CHARTER.md # Community decision-making processes
└── site/ # Project website (hub, demo, provocations, architecture)
See ROADMAP.md for the project phases and PLAN.md for the detailed development plan with work packages, timelines, and dependencies.
Phase 0 — Complete:
- WD-001 through WD-006: All 6 working documents (White Paper, Evidence Base, Signal Exchange Protocol, Explainer Script, Trust Fabric, Governance Framework)
- Protocol Integration Strategy (MCP/A2A/CIP standards path)
- Reference Implementation Outline (MVP pseudocode)
- Cooperative Taxonomy v0.1 (170+ entries with complementarity mappings)
- Interactive concept demo, 10 provocations, architecture diagram
- GitHub repository with issue templates and 11 seeded discussions
- 6 audience-specific outreach posts
- Unified project hub website
Phase 1 — In Progress (March-May 2026):
- MCP Spec Enhancement Proposal for
cooperative_contextresource type - A2A Agent Card extension for cooperative capability discovery
- CIP (Cooperative Intelligence Protocol) specification v0.1
- Working group formation (5 groups — join here)
- Pilot community identification (primary target: Newton, IL)
Phase 2 — Build (June-October 2026):
- MCP server reference implementation
- A2A cooperative discovery agents
- Community relay service
- Privacy audit and pilot deployment
We actively seek contributors from diverse disciplines. This is not a software project that happens to need some theory — it's a theoretical and design project that will eventually produce software. Right now, the most valuable contributions are ideas, critiques, and expertise.
- AI/ML — Context modeling, federated learning, agent architectures
- Cryptography — ZK proofs, MPC, homomorphic encryption, differential privacy
- Economics — Mechanism design, information economics, commons theory
- Network Science — Graph theory, diffusion dynamics, structural analysis
- Ethics — AI ethics, consent frameworks, power analysis, equity assessment
- Social Science — Trust formation, community dynamics, adoption research
- Systems Architecture — Protocol design, distributed systems, interoperability
- Domain Practitioners — Mutual aid, social work, community organizing
- Read the white paper in
docs/white-paper/ - Open a Discussion with your perspective, critique, or question
- Join a working group that matches your expertise
- Submit a PR with research, analysis, or specification work
- Challenge the assumptions — the most valuable contributions may be the ones that find flaws
See CONTRIBUTING.md for detailed guidelines.
| Group | Focus | Status |
|---|---|---|
| Theory & Foundations | Problem formalization, economic models, network theory | Forming |
| Privacy & Cryptography | Signal protocol, ZK matching, threat modeling | Forming |
| Ethics & Equity | Red-teaming, equity analysis, governance design | Forming |
| AI Integration | Context engines, MCP/A2A integration, agent design | Forming |
| Community & Practice | Practitioner input, pilot design, impact measurement | Forming |
Clarity about boundaries is as important as clarity about vision.
- Not a platform. Not a gig economy, social network, or marketplace. Those exist. This operates where they structurally cannot.
- Not an AI that decides. AI surfaces possibilities. Humans evaluate, choose, and act.
- Not a surveillance system. The architecture is designed so cooperation can happen without any central entity accessing private context.
- Not a utopian project. This addresses a real coordination failure through careful engineering. What people do with the infrastructure is up to them.
These works inform the theoretical foundations. This list will grow as research synthesis progresses.
Information Economics & Coordination
- Hayek, F.A. (1945). "The Use of Knowledge in Society"
- Hurwicz, L. (2008). "But Who Will Guard the Guardians?" (Nobel lecture on mechanism design)
Network Theory & Social Capital
- Granovetter, M. (1973). "The Strength of Weak Ties"
- Burt, R.S. (2004). "Structural Holes and Good Ideas"
- Putnam, R. (2000). Bowling Alone
Commons Governance
- Ostrom, E. (1990). Governing the Commons
- Ostrom, E. (2010). "Beyond Markets and States: Polycentric Governance"
Privacy-Preserving Computation
- Goldwasser, S., Micali, S., Rackoff, C. (1989). "The Knowledge Complexity of Interactive Proof Systems"
- Dwork, C. (2006). "Differential Privacy"
Cognitive Science of Expertise
- Kahneman, D. (2011). Thinking, Fast and Slow
- Ericsson, K.A. (2006). "The Influence of Experience and Deliberate Practice"
All foundational materials in this repository are released under Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0).
This means you are free to share, adapt, and build upon this work — even commercially — as long as you give appropriate credit and distribute derivative works under the same license.
Software implementations (when they exist) will use Apache 2.0.
This is a living project. It exists to be improved, challenged, and built upon.
If these ideas resonate, join us.