AI Safety
This page is maintained by Atlantic Cape Concepts LLC to explain how AI is used safely in SparkJournal.
Last updated: 7/21/2026
1. AI is optional
SparkJournal works without AI. Every free user can build vision boards, write reflection cards, and set goals using the built-in sticker and quote library. AI Studio is an optional paid feature for subscribers and adult accounts.
2. Curated presets for kid accounts
Accounts for users under 18 can only use pre-approved, curated AI style presets. Free-form custom prompts and uploads that bypass the preset library are not available to kid accounts. This reduces the chance of accidental or inappropriate generations.
3. What we block
We use automated input filters and output review to block content that attempts to generate:
- Sexual or sexually suggestive imagery, especially involving minors.
- Violence, gore, hate symbols, or harassment targeting real people.
- Content that promotes self-harm, eating disorders, or dangerous behavior.
- Personal information such as real phone numbers, addresses, or full names of non-public individuals.
4. Guardian visibility
When a verified parent or guardian links to a kid account, AI generations made by that account are visible to the guardian. Guardians can also pause AI features or pause the entire account from the Family Dashboard.
5. Human review
Flagged generations and repeated policy violations are reviewed by our team. Accounts that repeatedly attempt to bypass filters may be suspended or terminated. We may report content that indicates imminent harm to appropriate authorities.
6. Report something
If you see an AI-generated image that feels unsafe, wrong, or unexpected, use the report button inside the app or contact support. We investigate every report and remove content that violates our policies.
Not a certification
This page describes the controls and practices Atlantic Cape Concepts LLC has chosen to implement in SparkJournal. It is not an independent audit, certification, or legal guarantee. AI models are third-party services and outputs can be imperfect; we review and improve our filters on an ongoing basis.