AI Transparency
Last updated: December 2025
What's Changed
Added a -Security & Encryption (Not AI)- section clarifying that sensitive user-generated content in Journal, To-Do, Daily Anchors, and Flashcards is encrypted at rest and is not sent to the AI provider.
IkigaiApp now uses AI models to help turn your survey responses and taxonomy-based tags into readable, human-friendly insights. In particular, AI is used to generate your Ikigai Profile narrative, strengths, growth areas, and a practical action plan. We do not send your email, password, or payment details to the AI provider, and your data is not used to train models beyond what is necessary to provide the service.
Our AI features are designed as low-risk decision-support tools. They provide suggestions and reflections, not binding decisions, and you remain fully in control of how you use them.
- •Data minimization and purpose limitation
- •Bias mitigation and fairness
- •Explainability and user-friendly language
- •Human-in-the-loop for sensitive outcomes
AI processes a limited subset of your data: selected survey answers, Ikigai taxonomy tags (passions, skills, jobs) and derived profile scores. This is used only to generate narrative descriptions, strengths, growth suggestions, and action plans inside IkigaiApp. We do not send your email, login credentials, or payment details to the AI provider.
Separately from AI, some sensitive user-generated content (such as Journal entries, To-Do items, Daily Anchors tasks, and personal Flashcards) is encrypted at rest in the database. It is decrypted only after login to provide core features (such as search) and kept only in volatile memory during the active session. This content is not sent to the AI provider.
You can always ignore AI-generated suggestions, repeat the survey, or delete your account. You may also contact ai@ikigaiapp.life if you have questions, want human review, or wish to limit AI-driven processing where possible.
Some personalization features in IkigaiApp, such as similarity-based examples and Advanced Insights, use deterministic statistical methods (e.g. vector similarity and weighting) rather than artificial intelligence. These methods do not learn from user data and are fully explainable.