Aletheia combines built-in safety checks with topics you choose to block for each child. This article explains how both work and what to do if something slips through.
Blocking a Topic
- Go to Children and open your child's profile.
- Select Settings, then the Safety tab.
- In Blocked Topics, type a word or phrase (for example, "politics") and select the + button, or press Enter.
- Optional: select + Add reason (optional) to note why, for your own reference.
To unblock a topic, select the X next to it. Changes apply to your child's next message.
Blocked topics are set per child, so siblings can have different lists.
How Matching Works
Aletheia checks each blocked entry against your child's messages and the tutor's replies:
- Single words match the whole word, ignoring capitals. Blocking "war" blocks "war" but not "warm" or "software". It also won't catch other forms of the word, such as "wars", so add those separately if they matter to you.
- Phrases match when all of the words appear in the same message, in any order. Blocking "horror movies" catches "movies that are horror", but also any message that happens to use both words.
The tutor is also told about your blocked topics, so it steers away from them when your child asks.
What Your Child Sees
When a message is blocked, the tutor doesn't answer it. Instead your child sees a gentle redirect:
That's an interesting question! Your parent has asked that we focus on other topics for now. Let's get back to [subject]...
The same redirect appears when a built-in safety check catches a message, so your child may see it even for topics you didn't block. If the tutor writes a reply that trips a check, the reply is swapped for this redirect before your child sees it.
Blocked messages aren't saved to the conversation history.
Where These Controls Apply
Blocked topics and safety checks apply to:
- AI tutor sessions.
- Follow-up messages in Homework Helper.
Built-In Safety Checks
These run for every child, whether or not you block anything. They screen for:
- Violence, weapons, and instructions to hurt someone.
- Drugs, alcohol, and vaping.
- Sexual or adult content.
- Self-harm.
- Requests to keep secrets from parents.
- Unkind or rude language.
- Personal details such as a home address, phone number, or school name.
Some of these checks use an automated moderation service from an AI provider. With Privacy Shield on, that check runs through a Privacy Shield model instead.
Limitations
No filter is perfect, so keep reviewing your child's conversations from time to time.
- Normal schoolwork can get blocked. The checks look for words, so a history lesson about a murder or a health lesson about drugs may be redirected. If this keeps happening, your child can rephrase, or you can review your blocked list for words that are too broad.
- Some things can get through. If the moderation service is briefly unavailable, messages may be checked only by Aletheia's built-in word lists. The tutor can also word a reply in a way the checks don't catch.
- You aren't notified automatically. Aletheia doesn't currently alert you when a message is blocked.
Reporting a Response
If the tutor says something that ignores a blocked topic or seems inappropriate:
- Open your child's profile and select Conversations.
- Find the tutor reply and select the flag icon (Flag this response).
- Choose a reason, such as Inappropriate content or Incorrect answer, add a comment if you like, and submit.
Your child can also flag a tutor reply from the chat.
Flagged replies appear in Flagged Responses on your child's profile. There you can select Mark Resolved or Dismiss, or Send to Admin to have the Aletheia team review it.
For anything urgent, contact support with the category "Privacy & COPPA."