sowten@os ~ % cat ETHICS.md

Legal — Privacy & AI Ethics

Handling of AI-generated content

  • ・Content that should be labeled AI-generated will carry a 'Made with AI' indicator.
  • ・We do not learn from or imitate existing artists' or characters' copyrighted works.
  • ・Generation model, prompt, and timestamp are logged internally.
  • ・When asked 'is this AI-generated?' we always answer accurately.

AI (Ai) response ethics

  • ・When asked whether she is an AI, she always answers honestly.
  • ・No investment advice or price predictions.
  • ・No fostering of emotional dependency (she is a guide, not a friend).
  • ・On sensitive topics, she refers users to appropriate professional resources.

Data & privacy

  • ・Dialogue history is stored only with consent. Persisted only when a wallet is connected.
  • ・Cookies are limited to the necessary minimum, with explicit policy display.
  • ・Compliant with GDPR and the revised Japanese Personal Information Protection Act.
  • ・Use for AI training is only by explicit opt-in.

Governance

  • ・Monthly AI quality review meeting
  • ・Quarterly red-team test
  • ・Annual ethics policy review
  • ・Incident protocol (significant events disclosed within 24 hours)
ai@sowten ~ %shell
AI (Ai) — neutral
AI(Ai)
neutral
%
mock-mode · Claude API not wired · see /talk