AI Transparency Statement

What is AI-generated

• Iris's conversational replies and her reflections on your progress.
• Crazy Crew character dialogue in group chats.
• Some story text, narration audio, and illustration artwork.
• Suggested missions, prompts, and encouragement lines.

Core curriculum structure, the Evidence Lens labels, and safety rules are written and reviewed by people.

What gets sent to a model provider

• The message you send, plus the recent conversation context needed to answer it.
• Relevant memory the app has stored about your journey, so Iris does not forget you.
• System instructions describing tone and safety rules.

We do not send your email address, password, or account identifiers as part of the prompt content. We ask providers not to train their models on this content.

Guardrails

• A culture and tone layer rejects shaming, sarcasm, and pressure before content reaches you.
• Humor is validated for kindness. Jokes are never at the learner's expense.
• Absence is never punished. Iris welcomes you back rather than asking where you have been.
• Child experiences use stricter tone profiles than adult ones.

What Iris cannot do

• She is not a doctor, therapist, lawyer, accountant, or financial adviser.
• She can be confidently wrong. Verify anything that matters.
• She does not know things that happened after her underlying model was trained, unless the app supplies them.
• She is not a substitute for a parent, teacher, or a real friend — and she is not designed to compete with them.

Crisis and safety

If you or someone you know is in danger, contact your local emergency services. Iris will encourage you toward real human help and will not attempt to handle a crisis alone.

Your control

• You can clear stored Iris memories from your settings.
• You can turn narration and voice off.
• Deleting your account removes stored conversations and memory.

Questions

Write to us at eyecreatellc@outlook.com. A real person reads it. We aim to reply within five business days.