Skip to content

Science, governance & safety

Self-understanding requires more than a convincing answer.

The Human Manual is being designed so that memory, evidence, permissions, pattern confidence, safety rules, and progress are controlled by the application — not quietly invented by generated language.

Guiding principles

Six commitments.

01

Observation before explanation

02

Hypotheses before verdicts

03

Protection before pathology

04

Regulation before analysis

05

Evidence before pattern certainty

06

User correction before system confidence

The application should control

  • +Structured records
  • +Source provenance
  • +Pattern evidence
  • +Contradicting evidence
  • +Confidence state
  • +Permissions
  • +Sharing scope
  • +Retention
  • +Safety-mode transitions
  • +Progress records

Generated language may help with

  • +Explaining a structured observation
  • +Asking a guided question
  • +Summarizing user-confirmed content
  • +Offering alternative interpretations
  • +Rehearsing communication
  • +Translating structured information into readable language

Generated language must not control

  • —Clinical diagnosis
  • —Pattern truth
  • —Memory authority
  • —Permission decisions
  • —Crisis classification by itself
  • —Retention policy
  • —Sharing rights
  • —Clinical treatment decisions

Pattern confidence

Maturity, not certainty.

Every stage preserves both supporting and contradicting evidence.

  1. Stage 1

    Single Observation

  2. Stage 2

    Candidate

  3. Stage 3

    Emerging

  4. Stage 4

    User-Recognized

  5. Stage 5

    Recurring

  6. Stage 6

    Actively Changing

  7. Stage 7

    Outgrown

Safety limits

What the product is not.

  • —The Human Manual is not emergency care.
  • —It is not a substitute for a licensed professional.
  • —It should change behavior when urgent safety concerns arise.
  • —Safety pathways must be designed and reviewed before affected populations launch.
  • —Youth access requires a separate consent, privacy, safety, and retention architecture.

Governance roadmap

Intended structure.

  • +Clinical advisors
  • +Safety review
  • +Legal review
  • +Privacy review
  • +Product review
  • +Model evaluation
  • +Incident-response procedures
  • +Ongoing audit and correction

Note

Not all governance bodies are currently operational. This is a description of intended structure.

Understand how your data should be handled.