Docs·4ff474d·Updated Mar 14, 2026·43 ADRs
Back
ADR-020accepted

ADR-020: Trust-First Design Philosophy

ADR-020: Trust-First Design Philosophy

Date: 2025-12-29 Status: Accepted Deciders: Development Team Related: ADR-016 (Prestige), ADR-019 (Referral Trust)

Context

Platform design philosophy falls on a spectrum:

Zero-Trust Design (Current Platform Default):

  • Assume users will cheat/exploit
  • Design defensively around edge cases
  • Heavy monitoring, surveillance, verification
  • Creates hostile environment, discourages positive behavior
  • Examples: Uber's surveillance, Amazon's worker monitoring

Trust-First Design (Karmyq Approach):

  • Assume people are fundamentally good
  • Design for positive experiences
  • Light safeguards, heavy cultural reinforcement
  • Creates welcoming environment, encourages prosocial behavior
  • Examples: Wikipedia, Buy Nothing Project, cohousing communities

Core Question: Does focusing on cheaters create an environment that enforces cheating (self-fulfilling prophecy)?

Decision

Design Karmyq with trust-first philosophy: optimize for the 99%, not defend against the 1%.

Design Principles

1. People Are Fundamentally Good

  • Default assumption: users want to help and contribute
  • Platform strengthens this belief through positive reinforcement
  • Stories of successful exchanges highlighted
  • Gratitude and appreciation built into UX

2. Culture Over Code

  • Community norms more important than algorithmic rules
  • Restorative justice over punishment
  • Education over enforcement
  • Peer accountability over centralized moderation

3. Transparency Over Surveillance

  • Actions visible to community, not hidden algorithms
  • Clear consequences explained upfront
  • No black-box scoring or shadow banning
  • Members understand how trust works

4. Relationship Over Transaction

  • Every exchange builds (or damages) relationship
  • Long-term reputation more important than single interaction
  • Forgiveness and growth built into system
  • Second chances after mistakes

5. Local Governance

  • Communities set their own norms and standards
  • No universal "platform rules" imposed top-down
  • Cultural experimentation encouraged
  • Learn from each other's approaches

Trust Scaffolding (Not Walls)

Instead of preventing bad behavior, enable good behavior:

❌ Zero-Trust: Require ID verification, credit checks, background checks ✅ Trust-First: Referral chains, gradual privilege escalation, community vouching

❌ Zero-Trust: Ban users who get flagged ✅ Trust-First: Restorative conversations, mediation, community accountability

❌ Zero-Trust: Algorithmic risk scoring ✅ Trust-First: Visible karma and prestige earned through contribution

❌ Zero-Trust: Extensive rules and policies ✅ Trust-First: Simple norms co-created by community

❌ Zero-Trust: Centralized support tickets ✅ Trust-First: Peer-to-peer conflict resolution

When Trust Breaks Down

Acknowledge reality: Some people will behave badly. Handle it gracefully:

Restorative Justice Process:

  1. Private conversation between parties
  2. Mediation by trusted community member
  3. Community circle if unresolved
  4. Temporary suspension (not permanent ban)
  5. Path to reintegration after accountability

Community Protection:

  • Members can block others (local decision, not platform-wide)
  • Communities can vote to remove members (transparent process)
  • Extreme cases escalated to cross-community review
  • Focus on protecting vulnerable, not punishing offender

Learning from Incidents:

  • Anonymous incident reports for pattern recognition
  • Community discussions about norms
  • Cultural evolution, not rigid rules
  • Strengthen trust scaffold based on what breaks

Consequences

Positive

  • Welcoming Environment: New members feel trusted, not suspected
  • Prosocial Behavior: Trust encourages more trustworthy behavior
  • Cultural Coherence: Communities develop strong positive norms
  • Reduced Overhead: Less moderation, surveillance infrastructure
  • Authentic Relationships: Not transactional, performance-based

Negative

  • Vulnerable to Bad Actors: No hard barriers to entry (mitigated by referrals)
  • Slower Response: Restorative justice takes time vs instant bans
  • Emotional Labor: Conflict resolution falls on community, not platform
  • Cultural Risk: Weak communities may develop toxic norms
  • PR Risk: Single bad incident may damage platform reputation

Alternatives Considered

Alternative 1: Zero-Trust Design

  • Why rejected: Creates hostile environment; violates core mission of rebuilding trust

Alternative 2: Hybrid Model (Trust for Some, Surveillance for Others)

  • Why rejected: Creates two-tier system; still embeds surveillance infrastructure

Alternative 3: Algorithmic Moderation

  • Why rejected: Black box; removes human judgment; can't adapt to context

Implementation Notes

Phase 1: Cultural Foundation (v9.0)

  • Onboarding emphasizes trust-first philosophy
  • "People are good" messaging throughout UX
  • Positive story highlighting
  • Simple conflict resolution guidance

Phase 2: Restorative Justice Tools (v10.0)

  • Mediation request system
  • Community circle facilitation
  • Accountability tracking
  • Reintegration pathways

Phase 3: Pattern Learning (v11.0+)

  • Anonymous incident aggregation
  • Cross-community norm sharing
  • Cultural evolution metrics
  • Trust scaffold refinement

UX Examples

Welcoming Messages:

"Welcome to Portland Tools! This community runs on trust.
We believe you're here to help and be helped.
Your neighbor Sarah vouched for you - feel free to reach out if you have questions."

Request Form:

Instead of: "Report this user"
Use: "Start a conversation about this interaction"

Conflict Resolution:

Instead of: "Block and report"
Use: "Request mediation" or "Community circle"

Metrics to Track

Trust Indicators (what we measure):

  • Successful exchanges completed
  • Gratitude expressions
  • Repeat interactions
  • Community event attendance
  • Conflict resolution success rate

Not Measured (intentionally):

  • Surveillance metrics (clicks, time-on-site, engagement)
  • Punishment rates
  • Compliance scores
  • Conversion funnels

References

  • Zero-trust vs high-trust societies: Francis Fukuyama
  • Self-fulfilling prophecies: Robert Merton
  • Restorative justice: Howard Zehr
  • Gift economies: Lewis Hyde, "The Gift"
  • Wikipedia's trust model: Presumption of good faith
  • Buy Nothing Project: Community-driven norms