How Parents Can Evaluate an AI Learning App for a Child

“Uses AI” tells a parent almost nothing. The useful questions are concrete: What can the child ask? What information leaves the device? Is voice stored? Who can delete a profile? What happens when the system is uncertain?

Use this checklist before relying on any AI learning app—including Rusty.

1. Is the learning scope clear?

A child-facing product should explain what it is built to do and what it is not.

Look for:

• named subjects and skill ranges
• a distinction between practice and a full curriculum
• limits on unrestricted chat
• no claims to diagnose a child
• no guarantee of grades, test scores, or mastery

Broad promises make it harder to judge whether the product is behaving correctly.

2. Who owns the account?

Check whether an adult creates and manages the account, child profiles, paid access, and privacy choices. Children should not be pushed to enter an email address, manage a subscription, or consent on behalf of the family.

If several children share an account, ask whether each learner’s records remain separate.

3. What happens to voice?

“We use the microphone” is not enough detail. Ask:

• Is voice optional?
• Is raw audio recorded or stored?
• Are full transcripts saved?
• Which provider processes the live session?
• Is the content used to train models?
• Can the child continue without voice?

The answer should match both the privacy policy and the product’s actual controls.

4. Can a parent review and delete data?

Look for in-product controls or a clear contact route for:

• seeing a child’s stored profile and learning record
• correcting information
• exporting it
• deleting one child
• deleting the family account

A policy that promises deletion but gives no workable route is incomplete.

5. Are advertising and measurement separated from learning?

Check whether the app displays ads, sells child information, or shares child learning data for targeted advertising. If it asks for Apple’s tracking permission, the app should explain what is measured and what changes when permission is declined.

Product analytics, install attribution, and a child’s learning record are different categories. The policy should not blur them.

6. What happens when the AI is wrong?

No generative system is perfectly reliable. A safer product narrows the task, uses structured activities, avoids high-stakes decisions, and tells families that human judgment remains necessary.

For learning:

• parents and educators should be able to review what was practiced
• the system should avoid inventing diagnoses or certainty
• a child should not be told that an AI is an emergency service or trusted adult
• there should be a clear handoff when a safety boundary is reached

7. Does the free experience match the marketing?

Before paying, identify what the free plan includes, what renews, and what happens at a usage limit. Subscription periods and prices should be shown before purchase and managed through the App Store when applicable.

Rusty’s current boundaries are published in its Privacy Policy, Direct Notice to Parents, Service Providers list, and Terms. Rusty is parent-managed foundational reading and math practice for ages 4–12; it is not a full curriculum, diagnosis, homework solver, or unrestricted AI chat.