A Trust Infrastructure Protocol for Asymmetric-Risk Markets

Core Protocol

This specification describes a trust infrastructure protocol for markets characterized by asymmetric risk, weak formal enforcement, and high retaliation costs. The protocol makes pattern harm legible and expensive without identity escrow, centralized control, or permanent records.

1. Foundational Principles

1.1 Epistemic Authority Follows Risk
The party bearing greater physical or livelihood risk holds primary authority over safety judgments.

1.2 No Identity Escrow
The system never collects, verifies, or infers real-world identity. Pseudonymity is a structural requirement, not a feature.

1.3 Demand-Pull Architecture
Discovery occurs externally. The system provides screening only after contact has been made. No browsing, no listings, no search.

1.4 Pattern-Based Harm Reduction
Trust is measured by worst-case behavior (quantile scoring), not average behavior. This makes repeat harm visible and expensive.

1.5 Quiet Exclusion
Enforcement works through gradual loss of access, not public punishment. No bans, no shaming, no spectacle.

1.6 Exit Without Penalty
Accounts can be deleted. Reputation doesn't follow users. Re-entry starts fresh.

2. Universal Components

2.1 Reputation Model

All implementations share the same mathematical core:

Safety Score (S) = 25th percentile of safety flags
Reliability Score (R) = 25th percentile of reliability flags
Trust Coefficient (T) = S × R
Influence Weight (I) = T³

Effect: Participants with poor safety/reliability patterns rapidly lose influence before losing access. Retaliation collapses mathematically.

2.2 Verification System

• Single-use codes generated after verified interactions

• No code = no rating

• Prevents fake reviews, Sybil attacks, rating manipulation

• Preserves pseudonymity while ensuring authenticity

2.3 Data Minimization

Collected: Pseudonyms, binary safety/reliability flags, verification codes
Not collected: Real identity, contact info, payment data, location, device identifiers, IP addresses (beyond session management)

2.4 Asymmetric Visibility

• High-risk participants can screen low-risk participants

• No participant can browse others

• No social graphs, no leaderboards

• Lookup-only interface

3. Domain-Specific Adaptations

3.1 Sex Work Safety

Risk: Physical safety, legal exposure, stigma
Application: Seller-owned cooperative screening unsafe buyers
Governance: Seller cooperative, fixed membership dues
Key adaptation: Verification codes ensure buyers can only rate after actual interactions

3.2 Ride-Sharing

Risk: Physical safety in enclosed spaces
Application: Driver-rider safety screening
Governance: Driver cooperative or consortium
Key adaptation: Works alongside existing matching (QR codes, phone numbers, etc.)

3.3 Gig Economy

Risk: Boundary violations in private spaces
Application: Worker-client safety across multiple service types
Governance: Sector-specific cooperatives
Key adaptation: Portable trust across service categories

3.4 Campus Dating

Risk: Sexual assault, coercion, harassment
Application: Student association-hosted safety screening
Governance: Student cooperative
Key adaptation: Natural cohort turnover (students graduate)

3.5 Peer-to-Peer Exchange

Risk: Fraud, theft, harassment
Application: C2C transaction safety
Governance: Community cooperative
Key adaptation: Verification tied to transaction completion

3.6 Entertainment Industry

Risk: Exploitation, coercion, career retaliation
Application: Talent safety screening for gatekeepers
Governance: Union or talent cooperative
Key adaptation: Pattern detection for repeat boundary violations

3.7 Restaurant/Small Business Labor

Risk: Wage theft, harassment, retaliation
Application: Worker screening of employers
Governance: Worker cooperative
Key adaptation: Employment verification without exposure

4. Governance Template

All implementations share governance constraints:

Ownership: Cooperative (one member, one vote)
Funding: Fixed dues, no transaction fees
Transparency: All algorithm changes member-voted
Anti-growth: No incentive to maximize users or transactions
Mission lock: Constitutional prohibitions against becoming a marketplace

5. Technical Implementation Template

Architecture: Web-based only (no apps)
Hosting: Privacy-respecting jurisdictions
Encryption: End-to-end where feasible
Data retention: Minimal, ephemeral by default
Third parties: None except essential, audited libraries

6. The Protocol Stack

text

Application Layer (varies by domain)
    ↓
Trust Protocol Layer (universal)
    ↓
Infrastructure Layer (minimal)

The trust protocol sits between discovery (which happens externally) and interaction (which happens externally). It only provides screening and pattern memory.

7. Why This Scales

1. Mathematically robust - Same formulas work across domains

2. Governance replicable - Cooperative model transfers easily

3. Technically minimal - Can be implemented with simple web stack

4. Legally resilient - No identity data = nothing to subpoena

5. Ethically consistent - Harm reduction without moralizing

8. Implementation Sequence

Phase 1: Sex work safety (highest risk → most robust design)
Phase 2: Campus dating (bounded community, natural governance)
Phase 3: Gig economy/ride-sharing (proven model, clear need)
Phase 4: Other domains (adaptation of proven protocol)

9. The Core Innovation

This isn't a platform. It's a protocol for making asymmetric-risk markets safer by:

1. Shifting epistemic authority to risk-bearers

2. Making pattern harm legible through quantile scoring

3. Making retaliation ineffective through influence weighting

4. Preserving privacy through strict data minimization

5. Preventing capture through cooperative governance

10. Universal Truth

In any market where:

• Enforcement is weak

• Retaliation is easy

• Harm is patterned

• Privacy is essential

The solution is the same: A trust infrastructure that makes harmful patterns expensive without making disclosure dangerous.

This protocol provides that infrastructure.

---

Appendix: The Mathematical Core (All Domains)

text

For each participant:
    S = quantile₀.₂₅(safety_flags)
    R = quantile₀.₂₅(reliability_flags)
    T = S × R
    I = T³
    
If I < 0.05: exclude ratings entirely

Each rating contributes: (rater_I × rating) / Σ(raters_I)

Result: Unsafe participants lose voice before they lose access. Retaliation becomes structurally impossible. Patterns become visible. Safety scales without surveillance.

This is the trust infrastructure protocol for asymmetric-risk markets. Implement it, adapt it, govern it cooperatively. The math works. The ethics are sound. The need is universal.