AI Safety & Moriah

• Explain concepts in age-appropriate language and ask guiding questions.
• Offer hints, practice problems, examples, review prompts, and study support.
• Help a learner check reasoning, organize ideas, and improve a draft without taking ownership of the student’s work.
• Support parent-visible learning records and summaries when those features are enabled.

• Replace the parent, teacher of record, pastor, counselor, doctor, lawyer, or other human authority.
• Make autonomous high-impact educational decisions about a child.
• Complete graded student work when the assignment requires the student’s own response.
• Encourage secrecy from a parent or guardian, unrestricted web use, or evasion of safety controls.
• Diagnose a learner or present uncertain AI output as verified fact.

AI interactions that are stored for a family should be visible to the parent or guardian. Sensitive, unsafe, confusing, or high-stakes topics should be escalated to a responsible adult rather than treated as autonomous AI decision-making.

AI can produce inaccurate, incomplete, biased, or outdated content. Important facts, citations, calculations, and claims should be checked against reliable sources. Moriah should make uncertainty visible rather than inventing confidence.

Virtueward does not sell student data or use student activity for behavioral advertising. Identifiable student content is not intentionally submitted for general-purpose public model training. If a future AI feature materially changes what student information is processed, the parent-facing notice will be updated before the feature is enabled for student use.

Parents may report unsafe, inaccurate, inappropriate, or privacy-sensitive AI behavior to support@unshakenvoice.tech. Include the lesson or feature involved, but do not email passwords or unnecessary student-identifying material.